Antenna Positioning Analysis and Dual-Frequency Antenna ...€¦ · Antenna Positioning Analysis...

Antenna Positioning Analysis and

Dual-Frequency Antenna Design of

High Frequency Ratio for Advanced

Electronic Code Responding Labels

by

Kin Seong Leong

B.E. (Electrical & Electronic, First Class Honours),The University of Adelaide, Australia, 2003

Thesis submitted for the degree of

Doctor of Philosophy

in

The School of Electrical & Electronic Engineering,

The Faculty of Engineering, Computer and Mathematical Sciences,

The University of Adelaide, Australia

January 2008

c© 2008

Kin Seong Leong

All Rights Reserved

To

My Parents

Page iv

Contents

Contents v

Abstract xv

Statement of Originality xvii

Acknowledgment xix

Conventions xxi

Publications xxiii

List of Figures xxvii

List of Tables xxxvii

Chapter 1. Thesis Introduction 1

1.1 Thesis Title . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.2 Background of Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.3 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.4 Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1.5 Thesis Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

1.6 Chapter Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Chapter 2. RFID Systems 11

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.2 History of RFID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.3 A Simple RFID System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Page v

Contents

2.4 Variants of RFID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.4.1 Active and Passive RFID Tags . . . . . . . . . . . . . . . . . . . . . 14

2.4.2 Backscattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2.4.3 Common Frequency Bands for RFID Operations . . . . . . . . . . 16

2.5 RFID and Supply Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

2.6 The EPC Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

2.6.1 The Early Stage EPC Network . . . . . . . . . . . . . . . . . . . . 19

2.6.2 The EPC Network Current State . . . . . . . . . . . . . . . . . . . 21

2.7 RFID Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Chapter 3. Path Loss and Position Analysis 27

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

3.2 Path Loss Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

3.2.1 Free Space Path Loss . . . . . . . . . . . . . . . . . . . . . . . . . . 31

3.2.2 In-building Path Loss . . . . . . . . . . . . . . . . . . . . . . . . . 31

3.3 Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

3.4 Path Loss Model for RFID . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

3.5 Simple Two Reader Interference . . . . . . . . . . . . . . . . . . . . . . . . 34

3.6 Background for RFID Positioning Analysis . . . . . . . . . . . . . . . . . 37

3.6.1 Power Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

3.6.2 Antenna Gain Pattern . . . . . . . . . . . . . . . . . . . . . . . . . 39

3.6.3 Frequency Spectrum Channelling . . . . . . . . . . . . . . . . . . 40

3.6.4 Listen Before Talk . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

3.7 Simulation Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

3.7.1 Simulation Concepts Example . . . . . . . . . . . . . . . . . . . . 47

3.8 Results and Discussion (RITR) . . . . . . . . . . . . . . . . . . . . . . . . . 48

3.8.1 Modification of Existing Software . . . . . . . . . . . . . . . . . . 48

Page vi

Contents

3.8.2 Guidelines to Avoid RITR . . . . . . . . . . . . . . . . . . . . . . . 50

3.9 Results and Discussion (RRS) . . . . . . . . . . . . . . . . . . . . . . . . . 52

3.9.1 One Antenna Simulation . . . . . . . . . . . . . . . . . . . . . . . 53

3.9.2 Two Antenna Simulation . . . . . . . . . . . . . . . . . . . . . . . 55

3.9.3 One Antenna in Horizontal Position . . . . . . . . . . . . . . . . . 56

3.9.4 Antennas Operating in Different Channels . . . . . . . . . . . . . 57

3.9.5 A More Hostile Environment . . . . . . . . . . . . . . . . . . . . . 63

3.10 Real Life Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

3.10.1 A Checkout Counter . . . . . . . . . . . . . . . . . . . . . . . . . . 63

3.10.2 Reader Synchronisation . . . . . . . . . . . . . . . . . . . . . . . . 64

3.11 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

Chapter 4. RFID Operational Considerations 67

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

4.2 RFID EMC Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

4.2.1 Frequency Hopping Spread Spectrum (FHSS) . . . . . . . . . . . 70

4.2.2 Listen Before Talk (LBT) . . . . . . . . . . . . . . . . . . . . . . . . 70

4.3 RFID Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

4.4 Path Loss Measurement Experiment . . . . . . . . . . . . . . . . . . . . . 71

4.4.1 Experimental Setup: . . . . . . . . . . . . . . . . . . . . . . . . . . 72

4.4.2 Room Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

4.4.3 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . 73

4.4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

4.5 Sources of Simulation Error . . . . . . . . . . . . . . . . . . . . . . . . . . 80

4.5.1 Path Loss Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

4.5.2 Reflection, Refraction, and Diffraction . . . . . . . . . . . . . . . . 82

4.5.3 Radiation Pattern of Antenna . . . . . . . . . . . . . . . . . . . . . 84

Page vii

Contents

4.5.4 Simulation Result Interpretation and Analysis . . . . . . . . . . . 85

4.6 Second Carrier Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

4.6.1 Effect of the Position of Tags . . . . . . . . . . . . . . . . . . . . . 90

4.6.2 Effect of Environment Factor . . . . . . . . . . . . . . . . . . . . . 91

4.6.3 Combining First and Second Carrier Sensing . . . . . . . . . . . . 92

4.7 Investigation of Specific Absorption Rate (SAR) . . . . . . . . . . . . . . 96

4.7.1 SAR Level for UHF RFID Operation . . . . . . . . . . . . . . . . . 96

4.7.2 Experiment Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

4.7.3 Simulation of Power Density . . . . . . . . . . . . . . . . . . . . . 97

4.7.4 Maximum Power for Every Channel . . . . . . . . . . . . . . . . . 98

4.7.5 Worst-Case Scenario . . . . . . . . . . . . . . . . . . . . . . . . . . 100

4.7.6 Recommendations on SAR . . . . . . . . . . . . . . . . . . . . . . 102

4.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

Chapter 5. Reader Synchronisation 105

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

5.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

5.2.1 ETSI 302 208 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

5.2.2 EPC Class 1 Generation 2 Protocol . . . . . . . . . . . . . . . . . . 107

5.2.3 Problem in Dense Reader Environment . . . . . . . . . . . . . . . 108

5.3 Reader Synchronisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

5.4 Actual Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

5.4.1 Connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

5.4.2 Positioning of LBT Sensors . . . . . . . . . . . . . . . . . . . . . . 111

5.4.3 Antenna Positioning . . . . . . . . . . . . . . . . . . . . . . . . . . 113

5.5 Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

5.6 Synchronisation Fine-tuning . . . . . . . . . . . . . . . . . . . . . . . . . . 115

Page viii

Contents

5.6.1 Reduction of Output Power . . . . . . . . . . . . . . . . . . . . . . 116

5.6.2 Reduction of Overall Reader Talking Time . . . . . . . . . . . . . 117

5.6.3 Use of External Sensors . . . . . . . . . . . . . . . . . . . . . . . . 117

5.6.4 RF Opaque or RF Absorbing Materials . . . . . . . . . . . . . . . . 118

5.6.5 Frequent Rearrangement of Channels . . . . . . . . . . . . . . . . 118

5.7 Variation of Synchronisation . . . . . . . . . . . . . . . . . . . . . . . . . . 119

5.7.1 Separation of Transmitting and Receiving Channels . . . . . . . . 119

5.7.2 Separation of RFID and Non-RFID Signals . . . . . . . . . . . . . 121

5.8 Updated Progress on Development of RFID Reader Synchronisation . . 121

5.9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

Chapter 6. RFID Tag Antenna Design 123

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124

6.2 Antenna Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

6.3 Antenna Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

6.4 Resonant Circuit for Antenna . . . . . . . . . . . . . . . . . . . . . . . . . 137

6.5 Challenges in RFID Tag Antenna Design . . . . . . . . . . . . . . . . . . . 140

6.6 RFID Chips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144

6.7 RFID Readers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

6.7.1 HF Reader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

6.7.2 UHF Reader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

6.8 RFID Tag Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

6.9 Antenna Design and Simulation . . . . . . . . . . . . . . . . . . . . . . . 150

6.9.1 Ansoft HFSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

6.9.2 Scripting in HFSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

6.9.3 Plotting in HFSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

6.9.4 ISO-Pro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160

Page ix

Contents

6.10 HF RFID Antenna Design . . . . . . . . . . . . . . . . . . . . . . . . . . . 160

6.10.1 A Simple Loop Antenna . . . . . . . . . . . . . . . . . . . . . . . . 161

6.10.2 A HF Planar Spiral Coil Antenna . . . . . . . . . . . . . . . . . . . 162

6.10.3 An HF Antenna for a New Wine Closures . . . . . . . . . . . . . . 167

6.10.4 An HF Antenna for a Pigs . . . . . . . . . . . . . . . . . . . . . . . 168

6.11 UHF RFID Antenna Design . . . . . . . . . . . . . . . . . . . . . . . . . . 179

6.11.1 A Common UHF RFID Tag . . . . . . . . . . . . . . . . . . . . . . 181

6.11.2 Simple Planar Dipole . . . . . . . . . . . . . . . . . . . . . . . . . . 182

6.11.3 A UHF Antenna for Sheep . . . . . . . . . . . . . . . . . . . . . . . 183

6.11.4 A UHF Antenna for Pigs . . . . . . . . . . . . . . . . . . . . . . . . 185

6.11.5 A UHF Antenna for Beer Kegs . . . . . . . . . . . . . . . . . . . . 187

6.11.6 A UHF Antenna for New Wine Closures . . . . . . . . . . . . . . 188

6.12 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188

Chapter 7. Dual-Frequency RFID Antenna 191

7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192

7.2 Current Dual-Frequency Antenna Design . . . . . . . . . . . . . . . . . . 193

7.3 Design Aims . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194

7.4 Dual-Frequency Antenna Design . . . . . . . . . . . . . . . . . . . . . . . 196

7.4.1 Independent HF and UHF Antenna Design . . . . . . . . . . . . . 196

7.4.2 Quick Feasibility Test . . . . . . . . . . . . . . . . . . . . . . . . . . 198

7.4.3 Tunability Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200

7.4.4 Redesign of the UHF Dipole . . . . . . . . . . . . . . . . . . . . . 206

7.4.5 Compatibility of Dipole in HF . . . . . . . . . . . . . . . . . . . . 210

7.4.6 Merging of the New UHF Dipole with the HF Coil . . . . . . . . . 212

7.5 Antenna Fabrication and Testing . . . . . . . . . . . . . . . . . . . . . . . 217

7.6 Miniaturisation of Dual-Frequency RFID Antenna with High Frequency

Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

Page x

Contents

7.7 Improvement on Novel Dual-Frequency Antenna . . . . . . . . . . . . . 221

7.8 Testing on Miniaturised Antenna . . . . . . . . . . . . . . . . . . . . . . . 230

7.9 Final Design of Miniaturised Dual-Frequency Antenna . . . . . . . . . . 231

7.10 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231

Chapter 8. Alternative Dual-frequency Antenna Designs 235

8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236

8.2 Alternate Methods in Merging an HF Antenna and a UHF Antenna . . . 236

8.2.1 Different Feed Point Location . . . . . . . . . . . . . . . . . . . . . 237

8.2.2 Different UHF Antenna Types . . . . . . . . . . . . . . . . . . . . 239

8.3 HF Antenna Acting as UHF Antenna at UHF . . . . . . . . . . . . . . . . 241

8.3.1 Loss in Antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244

8.3.2 Antenna Measurement Results . . . . . . . . . . . . . . . . . . . . 245

8.4 UHF Antenna Acting as an HF Antenna at HF . . . . . . . . . . . . . . . 246

8.5 Multi-Feed Point Antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . 247

8.5.1 Stacking UHF and HF Antennas . . . . . . . . . . . . . . . . . . . 247

8.5.2 Side-by-side HF and UHF Antenna . . . . . . . . . . . . . . . . . 251

8.6 Comparison Between Different Methodologies . . . . . . . . . . . . . . . 251

8.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253

Chapter 9. Measurement Techniques 255

9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

9.2 Background and Literature Review . . . . . . . . . . . . . . . . . . . . . . 257

9.3 Settings and Connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259

9.4 Measurement Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262

9.4.1 A Balanced Bow Tie Antenna . . . . . . . . . . . . . . . . . . . . . 262

9.4.2 Half Bow Tie Antenna on Ground Plane . . . . . . . . . . . . . . . 264

9.4.3 Results Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . 264

Page xi

Contents

9.5 Discussion and Improvement . . . . . . . . . . . . . . . . . . . . . . . . . 265

9.6 Effect of Inaccuracy in Measurement . . . . . . . . . . . . . . . . . . . . . 268

9.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272

Chapter 10.Thesis Conclusions 273

10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274

10.2 Review of Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274

10.3 Possible Research Extensions . . . . . . . . . . . . . . . . . . . . . . . . . 275

10.4 Summary of Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . 276

10.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278

Appendix A. Calculation of Impedance of Coplanar Strip Line 279

Appendix B. Some Notes on the Concept of Inductance 281

Appendix C. MATLAB Code for Path Loss Calculation 285

C.1 Main Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285

C.2 Antenna Gain Pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286

C.3 Path Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288

Appendix D. MATLAB Code for HFSS VB Script Generation 291

Appendix E. MATLAB Code for Inductance Calculation 297

E.1 Main Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297

E.2 Inductance Calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

E.3 Mutual Inductance - Positive . . . . . . . . . . . . . . . . . . . . . . . . . 302

E.4 Mutual Inductance - Negative . . . . . . . . . . . . . . . . . . . . . . . . . 303

Appendix F. Path Loss Experiment 305

F.1 Preliminary Setting Up Procedure . . . . . . . . . . . . . . . . . . . . . . 305

F.2 Experiment Procedure and Results . . . . . . . . . . . . . . . . . . . . . . 306

Page xii

Contents

F.2.1 Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306

F.2.2 Signal Strength at Different Distances (Within and Between Build-

ings) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306

F.2.3 Reflection from Typical Wall and a Conductive Fence . . . . . . . 310

F.2.4 Propagation Loss Outdoors . . . . . . . . . . . . . . . . . . . . . . 312

F.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313

Appendix G. RFID Deployment in Piggery 315

G.1 Modelling of Pig Feeder . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315

G.2 Final Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317

G.2.1 Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . 317

G.2.2 Design of RFID Tag Antennas . . . . . . . . . . . . . . . . . . . . . 318

G.2.3 Feeder Design and Experiment Setup . . . . . . . . . . . . . . . . 322

G.2.4 Network Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322

G.2.5 Experiment Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 324

G.2.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329

Bibliography 333

Page xiii

Page xiv

Abstract

The research background of this thesis is Radio Frequency Identification (RFID), where

an object can be identified remotely using electromagnetic waves. The focus of this the-

sis is on the in-depth investigation of two major problems in the RFID deployment in

supply chain applications, namely the reader collision problem in dense reader envi-

ronments and the tag performance problem in hostile environments.

To resolve the reader collision problem, the first part of this thesis offers a compre-

hensive path loss model for the analysis of the positioning of RFID reader antennas.

Simulation software was developed to predict the signal strength at a certain distance

from a reader antenna in a dense reader environment.

This simulation software was also utilised to publish insights and research results in

four major areas, which are: (i) Investigation on the sources of error in RFID simula-

tion, to provide sensible and meaningful simulation results before actual deployment

of RFID readers. (ii) The development of the idea of reader synchronisation, mainly

to address the strict regulations imposed on the deployment of RFID readers in Eu-

rope. (iii) The determination of the threshold value for second carrier sensing in RFID,

to enable the proper enforcement of second carrier sensing to avoid tag confusion in

dense reader environments. (iv) The examination of Specific Absorption Rate (SAR) to

ensure human safety in a dense RFID reader environment.

The second part of this thesis addresses the RFID tag performance problem in hostile

environments. The focus is on the development of HF and UHF tags, from the initial

tag antenna design, tag antenna simulation, tag antenna prototyping and measure-

ment, to the manufacturing of fully functional RFID tags at laboratory standards by

combining RFID chips on to tag antennas.

Page xv

Abstract

Though there are existing commercial grade HF and UHF RFID tags, they are mostly

aimed at pallet level applications and are not suitable for deployment in hostile en-

vironments. The study cases presented in this thesis are mostly industrially driven,

where there is a need to design specialty HF and UHF tag antennas.

With a strong foundation in the development of HF and UHF RFID tags for various

industrially driven applications, the research then concentrates on the development

of a novel dual-frequency RFID antenna, which operates in both the HF and UHF

regions. This dual-frequency RFID tag antenna embraces the benefits of both the HF

and UHF tag antenna, which enable it to have a good read range while operating in

environments that pose difficulties for RFID technology, for example applications in

which ionised liquid is present, such as in cases of wine or bottled drinks.

Several methodologies were used to develop a dual-frequency antenna, including the

merging of HF and UHF antennas, and having a UHF resonance point on a typical HF

antenna. With the successful development of an original dual-frequency antenna, the

research was then expanded to miniaturise this dual-frequency antenna.

The benefits of RFID deployment in supply chains are undoubtedly massive, though

there are still issues and challenges to be resolved before a world-wide adoption is

possible. This thesis contributes in recommending various reader antenna positioning

and deployment techniques, and also contributes in developing HF tag antennas and

UHF tag antennas for hostile environments, and a novel dual-frequency tag antenna

to progress towards the aim of ubiquitous object identification.

Page xvi

Statement of Originality

This work contains no material that has been accepted for the award of any other de-

gree or diploma in any university or other tertiary institution and, to the best of my

knowledge and belief, contains no material previously published or written by an-

other person, except where due reference has been made in the text.

I give consent to this copy of my thesis being made available in the University Li-

brary.

The author acknowledges that copyright of published works contained within this

thesis (as listed in the publications page) resides with the copyright holder/s of those

works.

Signed Date

Page xvii

Page xviii

Acknowledgment

I would like to thank my principal supervisor Prof. Peter Cole for his guidance through-

out my research candidature. His constant encouragement and support have made the

completion of this thesis possible.

Also, I would like to acknowledge Prof. Bevan Bates and Dr. Samuel Mickan (my co-

supervisors) for their countless advice and assistance, and Alfio Grasso for his help in

arranging projects with industrial partners.

Many thanks to the administrative and technical support staff of School of Electrical

& Electronic Engineering, University of Adelaide, especially to Geoff Pook who fabri-

cated most of my prototype antennas.

Sincere thanks to anyone who has contributed in any technical discussion with me,

notably my fellow colleagues from Auto-ID Laboratory, and David Hall.

I am very grateful for my postgraduate studies financial supports, an Australian Post-

graduate Awards (APA) and the Frank Perry Scholarship. Also, special thanks to Pork

CRC for funding the RFID development and field trial in a piggery.

Also, I wish to extend my gratitude to industry partners who offered their help and

support in completing some of my projects, especially Dr. Ian McCauley from DPI

Victoria, and Bruce Dumbrell from Leader Products.

Last but not least, I really appreciate the unconditional love and support from my

family, and also from Mun Leng, my research partner and fiancee.

Kin Seong Leong

Dec. 2007

Page xix

Page xx

Conventions

Typesetting

This thesis is typeset in Times New Roman and Sans-Serif using LATEX2e. Referenc-

ing and citation style are based on the Institute of Electrical and Electronics Engineers

(IEEE) Transaction style [1]. For an electronic source, the last updated date of the source

is enclosed within round parentheses, and is placed immediately behind the author(s)’

name [2]. The last access date is included within square parentheses and can be found

at the end of the entry.

Units

The International System of Units (abbreviated SI units) [3] is used in this thesis. Pre-

fixes “nano”, “micro”, and “milli” are preferred but prefix “cm” is avoided.

Spelling

English spelling in this thesis is based on Australian English. One exception is in some

special cases where the proper noun is used. For example: “Auto-ID Center”, not

“Auto-ID Centre”. Also, the plural of “antenna” is chosen to be “antennas” not “an-

tennae”, to be in line with most of the international technical publications.

Page xxi

Page xxii

Publications

Book Chapter

[1] K. S. Leong, M. L. Ng, and P. H. Cole, “RFID reader synchronisation,” in RFID Handbook: Applica-

tions, Technology, Security, and Privacy, S. Ahson and M. Ilyas, Eds. CRC, 2008.

[2] M. L. Ng, K. S. Leong, and P. H. Cole, “RFID tags for metallic object identification,” in RFID Hand-

book: Applications, Technology, Security, and Privacy, S. Ahson and M. Ilyas, Eds. CRC, 2008.

Journal

[1] K. S. Leong, M. L. Ng, A. Grasso, and P. H. Cole, “Dense RFID reader deployment in Europe using

synchronization,” Journal of Communications, vol. 1, no. 7, pp. 9–16, 2006.

Conference

[1] K. S. Leong, M. L. Ng, and P. H. Cole, “HF and UHF RFID tag design for pig tagging,” in 11th

Biennial Conference of the Australasian Pig Science Association (APSA), Brisbane, Australia, 25-28 Nov.

2007.

[2] K. S. Leong, M. L. Ng, and P. H. Cole, “Investigation on the deployment of HF and UHF RFID

tag in livestock identification,” in IEEE Antennas and Propagation Society International Symposium,

Honolulu, Hawaii, USA, 10-15 Jun. 2007.

[3] K. S. Leong, M. L. Ng, and P. H. Cole, “Miniaturization of dual-frequency RFID antenna with

high frequency ratio,” in IEEE Antennas and Propagation Society International Symposium, Honolulu,

Hawaii, USA, 10-15 Jun. 2007.

[4] K. S. Leong, M. L. Ng, and P. H. Cole, “Investigation of RF cable effect on RFID tag antenna

impedance measurement,” in IEEE Antennas and Propagation Society International Symposium, Hon-

olulu, Hawaii, USA, 10-15 Jun. 2007.

[5] K. S. Leong, M. L. Ng, and P. H. Cole, “Investigation of the threshold of second carrier sensing in

RFID deployment,” in 2006 International Symposium on Applications and the Internet (SAINT) Work-

shop, RFID and Extended Network: Deployment of Technologies and Applications, Hiroshima, Japan,

15-19 Jan. 2007.

Page xxiii

Publications

[6] K. S. Leong, M. L. Ng, and P. H. Cole, “Dual-frequency antenna design for RFID application,” in

21st International Technical Conference on Circuits/Systems, Computers and Communications (ITC-CSCC

2006), Chiang Mai, Thailand, 10-13 July 2006.

[7] K. S. Leong, M. L. Ng, and P. H. Cole, “Operational considerations in simulation and deployment of

RFID systems,” in 17th International Zurich Symposium on Electromagnetic Compatibility, Singapore,

27 Feb. - 3 Mar. 2006.

[8] K. S. Leong, M. L. Ng, A. Grasso, and P. H. Cole, “Synchronisation of RFID readers for dense

RFID reader environments,” in 2006 International Symposium on Applications and the Internet (SAINT)

Workshop, RFID and Extended Network: Deployment of Technologies and Applications, Phoenix, Arizona,

USA, 23-27 Jan. 2006.

[9] K. S. Leong, M. L. Ng, and P. H. Cole, “Positioning analysis of multiple antennas in a dense RFID

reader environment,” in 2006 International Symposium on Applications and the Internet (SAINT) Work-

shop, RFID and Extended Network: Deployment of Technologies and Applications, Phoenix, Arizona,

USA, 23-27 Jan. 2006.

[10] K. S. Leong, M. L. Ng, and P. H. Cole, “The reader collision problem in RFID systems,” in IEEE

2005 International Symposium on Microwave, Antenna, Propagation and EMC Technologies for Wireless

Communications, Beijing, China, 8-12 Aug. 2005.

[11] M. L. Ng, K. S. Leong, and P. H. Cole, “Design and miniaturization of an RFID tag using a sim-

ple rectangular patch antenna for metallic object identification,” in IEEE Antennas and Propagation

Society International Symposium, Honolulu, Hawaii, USA, 10-15 Jun. 2007.

[12] M. L. Ng, K. S. Leong, and P. H. Cole, “A small passive UHF RFID tag for metallic item identifi-

cation,” in 21st International Technical Conference on Circuits/Systems, Computers and Communications

(ITC-CSCC 2006), Chiang Mai, Thailand, 10-13 July 2006.

[13] M. L. Ng, K. S. Leong, D. M. Hall, and P. H. Cole, “A small passive UHF RFID tag for livestock

identification,” in IEEE 2005 International Symposium on Microwave, Antenna, Propagation and EMC

Technologies for Wireless Communications, Beijing, China, 8-12 Aug. 2005.

[14] M. L. Ng, K. S. Leong, and P. H. Cole, “Analysis of constraints in small UHF RFID tag design,”

in IEEE 2005 International Symposium on Microwave, Antenna, Propagation and EMC Technologies for

Wireless Communications, Beijing , China, 8-12 Aug. 2005.

[15] D. C. Ranasinghe, K. S. Leong, M. L. Ng, D. W. Engels, and P. H. Cole, “A distributed architecture

for a ubiquitous RFID sensing network,” in 2nd International Conference on Intelligent Sensors, Sensor

Networks and Information Processing (ISSNIP), Melbourne, Australia, 5-8 Dec. 2005.

[16] D. C. Ranasinghe, K. S. Leong, M. L. Ng, D. W. Engels, and P. H. Cole, “A distributed architec-

ture for a ubiquitous item identification network,” in Seventh International Conference on Ubiquitous

computing, Tokyo, Japan, 11-14 Sept. 2005.

Page xxiv

Publications

Non-refereed

[1] K. S. Leong, and M. L. Ng, “A simple EPC enterprise model,” in Auto-ID Labs Workshop, Zurich, 23-24

Sept. 2004.

[2] K. S. Leong, M. L. Ng, and D. W. Engels, “EPC network architecture,” in Auto-ID Labs Workshop,

Zurich, 23-24 Sept. 2004.

[3] M. L. Ng, K. S. Leong, and D. W. Engels, “Prospects for ubiquitous item identification,” in Auto-ID

Labs Workshop, Zurich, 23-24 Sept. 2004.

Page xxv

Page xxvi

List of Figures

1.1 A simple RFID system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.2 The complete structure of the thesis . . . . . . . . . . . . . . . . . . . . . 7

2.1 RFID system: reader and tag relationship . . . . . . . . . . . . . . . . . . 14

2.2 The principle of backscattering . . . . . . . . . . . . . . . . . . . . . . . . 15

2.3 Early stage EPC Network . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2.4 The Current EPC Network . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

3.1 Experimental path loss results . . . . . . . . . . . . . . . . . . . . . . . . . 33

3.2 Plot of proposed piece-wise linear in-building path loss model against

distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

3.3 A simple illustration of Reader Interference with Tag Replies (RITR) . . 35

3.4 Polar plot of the antenna gain of a directional circularly polarised RFID

antenna with a gain of 6 dBi . . . . . . . . . . . . . . . . . . . . . . . . . . 39

3.5 Transmit mask for multiple-interrogator and dense-interrogator envi-

ronments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

3.6 Grid used in simulation software . . . . . . . . . . . . . . . . . . . . . . . 45

3.7 The orientation and the position of a transmitting antenna with respect

to the simulation grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

3.8 Simulation results verifying the functionality of the software developed

in MATLAB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

3.9 Common positionings of antennas with respect to each other . . . . . . . 51

3.10 Simulation results on one antenna simulation . . . . . . . . . . . . . . . . 53

3.11 Simulation results of a two antenna simulation . . . . . . . . . . . . . . . 56

Page xxvii

List of Figures

3.12 The configuration of a horizontal antenna on ground or facing the ground

with respect to the normal configuration of a transmitting antenna . . . 57

3.13 Simulation results showing the effect of horizontal antenna configuration 58

3.14 Simulation results for some antennas operating in different channels and

pointing in different directions . . . . . . . . . . . . . . . . . . . . . . . . 60

3.15 Determination of safe distance in accordance to LBT when the second

antenna is an isotropic radiator . . . . . . . . . . . . . . . . . . . . . . . . 61

3.16 Repositioning of a horizontal antenna . . . . . . . . . . . . . . . . . . . . 64

3.17 Reader synchronisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

4.1 Equipment used in path loss prediction experiment . . . . . . . . . . . . 73

4.2 Experiment grid for signal strength measurement . . . . . . . . . . . . . 74

4.3 Measuring configuration from side view . . . . . . . . . . . . . . . . . . . 78

4.4 Plot of path loss using different models together with measured results. 81

4.5 Comparison of the received signal strength with and without the pres-

ence of a nearby metallic box . . . . . . . . . . . . . . . . . . . . . . . . . 83

4.6 Polar plot of the antenna gain of a directional circularly polarised RFID

antenna with a gain of 6 dBi . . . . . . . . . . . . . . . . . . . . . . . . . . 85

4.7 Boundary zone in simulation results . . . . . . . . . . . . . . . . . . . . . 86

4.8 Second carrier sensing with “Listen Before Talk” provision . . . . . . . . 88

4.9 Common antenna positioning . . . . . . . . . . . . . . . . . . . . . . . . . 89

4.10 Arrangement of 64 reader antennas for SAR simulation . . . . . . . . . . 98

4.11 The experiment site of the 64 reader antenna experiment . . . . . . . . . 99

4.12 Plot of simulation results on SAR investigation . . . . . . . . . . . . . . . 102

5.1 Synchronisation of all readers . . . . . . . . . . . . . . . . . . . . . . . . . 110

5.2 Centralised LBT system . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

Page xxviii

List of Figures

5.3 Localised LBT system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

5.4 Antenna positioning in dock door . . . . . . . . . . . . . . . . . . . . . . 113

5.5 Alternating of “Listening” and “Talking” mode . . . . . . . . . . . . . . . 115

5.6 Channelling of the allocated frequency spectrum . . . . . . . . . . . . . . 115

5.7 Reduction of output power to fine-tune reader synchronisation . . . . . 116

5.8 Using sensors in an RFID system . . . . . . . . . . . . . . . . . . . . . . . 117

5.9 Use of RF absorbing materials . . . . . . . . . . . . . . . . . . . . . . . . . 118

5.10 Channel switching within antennas . . . . . . . . . . . . . . . . . . . . . . 119

5.11 The complete frequency band allocated for RFID operation in Europe . . 120

5.12 Separation of transmitting and receiving channels in the frequency band

allocated for RFID operation in Europe . . . . . . . . . . . . . . . . . . . 120

6.1 The difference between source and vortex . . . . . . . . . . . . . . . . . . 126

6.2 Boundary conditions affecting electric and time varying magnetic field . 127

6.3 Spherical coordinate system . . . . . . . . . . . . . . . . . . . . . . . . . . 127

6.4 Series resonant circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

6.5 Parallel resonant circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

6.6 Practical parallel resonant circuit . . . . . . . . . . . . . . . . . . . . . . . 139

6.7 Reflection coefficient for best utilisation of λRC . . . . . . . . . . . . . . . . 142

6.8 The proposed chip bonding on to a thin metal strip . . . . . . . . . . . . 144

6.9 Reuse of HF RFID tag chip . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

6.10 A ISD72128 or C220 HF RFID tag . . . . . . . . . . . . . . . . . . . . . . . 145

6.11 72128 TTF chip on smart card module . . . . . . . . . . . . . . . . . . . . 146

6.12 A Texas Instruments UHF strap . . . . . . . . . . . . . . . . . . . . . . . . 146

6.13 An Alien C1G2 UHF strap . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

6.14 FEIG HF RFID Reader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

6.15 Gemplus HF RFID reader . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

Page xxix

List of Figures

6.16 FEIG UHF RFID reader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

6.17 Alien UHF RFID reader . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

6.18 Typical HF antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150

6.19 Effect of using Z-axis conductive tape on an HF RFID tag . . . . . . . . . 151

6.20 PML layer in HFSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

6.21 Linking MATLAB and HFSS . . . . . . . . . . . . . . . . . . . . . . . . . . 156

6.22 The auto generated loop using VBS . . . . . . . . . . . . . . . . . . . . . . 157

6.23 Plots |S11| curve in HFSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

6.24 Plots in HFSS showing the normalised |S11| curve . . . . . . . . . . . . . 159

6.25 Computation of the mutual inductance between two planar wire segments165

6.26 A Zork wine closure model . . . . . . . . . . . . . . . . . . . . . . . . . . 167

6.27 RFID tag for wine closure . . . . . . . . . . . . . . . . . . . . . . . . . . . 167

6.28 Casing for livestock RFID tag . . . . . . . . . . . . . . . . . . . . . . . . . 170

6.29 Tag encapsulation casing from Leader Products . . . . . . . . . . . . . . . 170

6.30 First version of a pig tag with bent tracks to connect all the circular tracks

together. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

6.31 Second version of a pig tag using a spiral track . . . . . . . . . . . . . . . 171

6.32 First prototype of a pig tag . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

6.33 Problem of uneven track thickness when prototyping antenna with thin

tracks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

6.34 First version prototype of a pig tag . . . . . . . . . . . . . . . . . . . . . . 173

6.35 Pig tag with TTF compatibility . . . . . . . . . . . . . . . . . . . . . . . . 174

6.36 Testing a pig tag with TTF compatibility . . . . . . . . . . . . . . . . . . . 174

6.37 Pig tag with external capacitance added . . . . . . . . . . . . . . . . . . . 176

6.38 Simulation results for the pig tag in Fig. 6.37 at HF with a thickness of

1.6 mm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176

6.39 Simulation results for the pig tag in Fig. 6.37 at HF with a thickness of

0.8 mm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177

Page xxx

List of Figures

6.40 Final version prototype of a pig tag . . . . . . . . . . . . . . . . . . . . . . 177

6.41 HF RFID Tag to be embedded in a livestock ear tag . . . . . . . . . . . . 178

6.42 RFID tags before and after encapsulation process . . . . . . . . . . . . . . 179

6.43 An encapsulated tag immersed in water to test for readability and water

proof capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179

6.44 Overall RFID system setup in piggery . . . . . . . . . . . . . . . . . . . . 180

6.45 HF and UHF RFID readers in piggery . . . . . . . . . . . . . . . . . . . . 180

6.46 Orientation of HF and UHF tags on pigs ears . . . . . . . . . . . . . . . . 180

6.47 Picture on pigs feeding in an RFID pig feeder . . . . . . . . . . . . . . . . 181

6.48 Sheep ear tag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184

6.49 Two UHF pig tags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185

6.50 Structure of the RFID tag with a rectangular loop antenna . . . . . . . . . 187

6.51 RFID tag for wine closure . . . . . . . . . . . . . . . . . . . . . . . . . . . 188

7.1 A generic HF coil antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . 197

7.2 A generic UHF dipole antenna . . . . . . . . . . . . . . . . . . . . . . . . 197

7.3 First version of a dual-frequency antenna . . . . . . . . . . . . . . . . . . 199

7.4 Impedance of the first version of dual-frequency antenna at HF . . . . . 199

7.5 Impedance of the first version of dual-frequency antenna at UHF . . . . 200

7.6 Second version of dual-frequency antenna . . . . . . . . . . . . . . . . . . 201

7.7 Impedance of the second version of dual-frequency antenna at UHF . . 202

7.8 Graph showing the calculated impedance and simulated impedance of

the second version of dual-frequency antenna . . . . . . . . . . . . . . . . 205

7.9 Location of feed point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205

7.10 Smith Chart showing the simulated impedance of the dual-frequency

antenna with different transmission line length . . . . . . . . . . . . . . . 207

7.11 A UHF dipole with a matching network . . . . . . . . . . . . . . . . . . . 208

Page xxxi

List of Figures

7.12 A UHF dipole with tuned matching network . . . . . . . . . . . . . . . . 208

7.13 Impedance of the UHF dipole with tuned matching network . . . . . . . 209

7.14 A UHF dipole with bent sides . . . . . . . . . . . . . . . . . . . . . . . . . 210

7.15 Impedance of the UHF dipole with bent sides . . . . . . . . . . . . . . . . 210

7.16 A UHF dipole with bent sides and gap . . . . . . . . . . . . . . . . . . . . 211

7.17 A UHF dipole with bent sides and repositioned gap . . . . . . . . . . . . 211

7.18 Final version of the UHF dipole, with matching network, bent sides and

gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212

7.19 HF Impedance of the final version of the UHF dipole shown in Fig 7.18 . 213

7.20 UHF Impedance of the final version of the UHF dipole shown in Fig 7.18 213

7.21 Self resonance frequency of HF planar coil antenna shown in Fig. 7.6 . . 214

7.22 Merging of HF multi-turn loop antenna with a UHF dipole . . . . . . . . 215

7.23 Final design for dual-frequency antenna . . . . . . . . . . . . . . . . . . . 216

7.24 Simulated impedance for the dual-frequency antenna at HF . . . . . . . 217

7.25 A transmission measurement using a wide-band loop antenna . . . . . . 218

7.26 Fabricated final version of the designed dual-frequency antenna . . . . . 218

7.27 Transmission measurement using network analyser to locate the reso-

nance point at HF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219

7.28 The simulated and measured results for the frequency response of the

fabricated dual-frequency antenna within the RFID UHF band from 860

- 960 MHz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

7.29 A different version of a dual-frequency antenna where the dipole is lo-

cated in the inside of the HF coil . . . . . . . . . . . . . . . . . . . . . . . 222

7.30 The HF impedance of the dual-frequency antenna in Fig 7.29 . . . . . . . 222

7.31 The UHF impedance of the dual-frequency antenna in Fig 7.29 . . . . . . 223

7.32 A miniaturised HF coil antenna . . . . . . . . . . . . . . . . . . . . . . . . 223

7.33 A miniaturised UHF dipole . . . . . . . . . . . . . . . . . . . . . . . . . . 224

7.34 A dual-frequency antenna with miniaturised HF coil and UHF dipole . . 225

Page xxxii

List of Figures

7.35 Impedance transformation in Smith Chart where all the traces cover 860

to 960 MHz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226

7.36 Dual-frequency antenna with rotated HF coil . . . . . . . . . . . . . . . . 228

7.37 Simulated results at HF for antenna in Fig. 7.36 . . . . . . . . . . . . . . . 229

7.38 Simulated results at UHF for antenna in Fig. 7.36 . . . . . . . . . . . . . . 229

7.39 Figure to illustrate the position of RFID chip attached to the antenna . . 230

7.40 A miniaturised dual-frequency antenna . . . . . . . . . . . . . . . . . . . 232

7.41 Final antenna design with design parameters . . . . . . . . . . . . . . . . 232

8.1 Feed point repositioning of the dual-frequency antenna . . . . . . . . . . 237

8.2 Impedance of dual-frequency antenna with feed point at position A at HF238

8.3 Impedance of dual-frequency antenna with feed point at location A at

UHF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238

8.4 A UHF patch antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240

8.5 The HF antenna to be fine-tuned to operate at UHF . . . . . . . . . . . . 242

8.6 Simulated impedance of the HF antenna shown in Fig. 8.5 at HF . . . . . 242

8.7 Simulated impedance of the HF antenna shown in Fig. 8.5 at UHF . . . . 243

8.8 The structure of the HF antenna shown in Fig. 8.5 with illustration of

surface current density on the top side . . . . . . . . . . . . . . . . . . . . 244

8.9 Measurement result of the HF antenna shown in Fig. 8.5 at UHF . . . . . 246

8.10 Stacking of HF and UHF RFID tags . . . . . . . . . . . . . . . . . . . . . . 248

8.11 Frequency response of the HF tag at UHF without the presence of any

UHF tag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

8.12 Frequency response of the UHF tag at UHF when placed near to HF tag 250

8.13 Frequency response of the UHF tag at UHF when the tag is able to be

read by an interrogator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250

Page xxxiii

List of Figures

9.1 Explanation of common mode current . . . . . . . . . . . . . . . . . . . . 258

9.2 Cables used for measuring experiment . . . . . . . . . . . . . . . . . . . . 260

9.3 A bow tie antenna on FR4 . . . . . . . . . . . . . . . . . . . . . . . . . . . 260

9.4 Half bow tie antenna on ground plane . . . . . . . . . . . . . . . . . . . . 261

9.5 The effect of SMA connector on measurement results . . . . . . . . . . . 262

9.6 Simulated and measured resistance of AUT . . . . . . . . . . . . . . . . . 263

9.7 Simulated and measured reactance of AUT . . . . . . . . . . . . . . . . . 263

9.8 The designed Balun . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267

9.9 The power transfer efficiency between AUT and chip when reactances

are equal and opposite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270

9.10 The power transfer efficiency between AUT and chip when resistances

are equal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271

9.11 The comparison between the estimated and calculated values on the

power loss curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271

B.1 Illustration to explain the calculation of flux density of a round wire of

radius a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283

F.1 Antennas on tripods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305

F.2 Measurement of signal strength . . . . . . . . . . . . . . . . . . . . . . . . 306

F.3 Path taken when measuring signal strength from the west side of Engi-

neering North building . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307

F.4 Path taken when measuring signal strength from the east side of Engi-

neering North building . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307

F.5 Path taken when measuring signal strength from Engineering and Maths

(EM) building . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307

F.6 Comparison between calculated and measured received signal strength

(in building) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310

Page xxxiv

List of Figures

F.7 Antennas facing directly perpendicular to the wall . . . . . . . . . . . . . 311

F.8 Antennas facing with an angle to the wall . . . . . . . . . . . . . . . . . . 311

F.9 Antennas on tripods facing a conductive fence . . . . . . . . . . . . . . . 311

F.10 Measurement of signal strength in outdoors . . . . . . . . . . . . . . . . . 313

F.11 Antennas on the ground and facing upwards . . . . . . . . . . . . . . . . 313

G.1 A pig feeder used in RFID deployment experiment . . . . . . . . . . . . 316

G.2 Modelling of a pig feeder (Stage 1) . . . . . . . . . . . . . . . . . . . . . . 317

G.3 Modelling of a pig feeder (Stage 2) . . . . . . . . . . . . . . . . . . . . . . 318

G.4 Overview arrangement of RFID deployment in piggery . . . . . . . . . . 323

G.5 The position of an antenna in an antenna box besides a pig feeder . . . . 324

G.6 Encapsulation of a HF reader . . . . . . . . . . . . . . . . . . . . . . . . . 325

G.7 Encapsulation of a UHF reader . . . . . . . . . . . . . . . . . . . . . . . . 326

G.8 Network setup in a piggery . . . . . . . . . . . . . . . . . . . . . . . . . . 326

Page xxxv

Page xxxvi

List of Tables

3.1 The effect of tag distance on multi-reader interference . . . . . . . . . . . 36

3.2 UHF RFID reader radiated power with corresponding threshold values

for LBT, and minimum allowable distance between antennas . . . . . . . 42

3.3 Safe distances, in m, for different antenna positioning to avoid the Reader

Interference with Tag Replies (RITR) problem . . . . . . . . . . . . . . . . 52

3.4 Safe distance for different antenna positionings to avoid Regulatory Reader

Shutdown (RRS) problem . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

3.5 Safe distance for different antenna configuration when the second an-

tenna is an isotropic radiator in accordance to LBT . . . . . . . . . . . . . 62

4.1 UHF RFID reader effective radiated power with corresponding thresh-

old values for LBT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

4.2 Average prediction of path loss based on experimental results . . . . . . 75

4.3 Difference in wavelength between direct distance, r, and reflected path, D 79

4.4 Comparison of measured and predicted signal strength . . . . . . . . . . 84

4.5 Threshold values for second carrier sensing with respect to tag read

range for different antenna orientations . . . . . . . . . . . . . . . . . . . 92

4.6 Threshold values for second carrier sensing with respect to environment

factor, n, for different antenna orientations . . . . . . . . . . . . . . . . . 93

4.7 Simulation results on SAR investigation using maximum power for ev-

ery channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

4.8 Calculation results of the worst-case scenario . . . . . . . . . . . . . . . . 101

5.1 UHF RFID reader radiated power with corresponding threshold values

for LBT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

5.2 Safe distance for different antenna configuration when the second an-

tenna is an isotropic radiator in accordance to LBT . . . . . . . . . . . . . 109

Page xxxvii

List of Tables

6.1 SI units for common terms used in antenna theory . . . . . . . . . . . . . 125

6.2 Inductance values for spiral planar coil with different numbers of turns,

with L = 20; L1=40; w=3; g=1 . . . . . . . . . . . . . . . . . . . . . . . . . 157

6.3 Inductance values for various spiral planar coils . . . . . . . . . . . . . . 166

6.4 Performance of HF RFID tag for Zork . . . . . . . . . . . . . . . . . . . . 168

6.5 Read range for UHF tags . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186

6.6 Read range for UHF tag for beer keg . . . . . . . . . . . . . . . . . . . . . 187

6.7 Performance of UHF RFID tag for the Zork wine closure . . . . . . . . . 188

7.1 HF versus UHF in RFID operation . . . . . . . . . . . . . . . . . . . . . . 193

7.2 Calculation of combined impedance based on individual impedances . . 203

7.3 Comparison between calculated impedances and simulated impedances 204

8.1 Measurement results for the UHF read range of the dual-frequency an-

tenna using the stacking method . . . . . . . . . . . . . . . . . . . . . . . 249

8.2 Comparison between different methodologies in creating a dual-frequency

RFID tag antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252

9.1 Average error and maximum error . . . . . . . . . . . . . . . . . . . . . . 266

F.1 Measured and calculated signal strengths for various locations . . . . . . 309

F.2 Measured signal strength for antennas facing a wall or a conductive fence 312

F.3 Calculated and measured signal strength (outdoor) for different antenna

orientations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314

G.1 Performance of HF RFID tag before and after encapsulation . . . . . . . 320

G.2 Performance of UHF RFID tag before and after encapsulation . . . . . . 321

G.3 HF and UHF RFID tags attached to pigs . . . . . . . . . . . . . . . . . . . 327

G.4 Results from remote monitoring . . . . . . . . . . . . . . . . . . . . . . . . 330

Page xxxviii

Chapter 1

Thesis Introduction

THIS chapter explains the title of this thesis, which is “Antenna

Positioning Analysis and Dual-Frequency Antenna Design of

High Frequency Ratio for Advanced Electronic Code Responding

Labels”. The background of the research topic, Radio Frequency Identifica-

tion (RFID), is presented. The motivation and contribution to this research

is discussed, followed by the elaboration of the structure of this thesis.

Page 1

1.1 Thesis Title

1.1 Thesis Title

The title of this thesis is “Antenna Positioning Analysis and Dual-Frequency Antenna

Design of High Frequency Ratio for Advanced Electronic Code Responding Labels”.

There are two related areas of research presented in this thesis. The first area is “An-

tenna Positioning Analysis”, while the second area is “Dual-Frequency Antenna De-

sign of High Frequency Ratio for Advanced Electronic Code Responding Labels”.

In “Antenna Positioning Analysis”, the term “Antenna” refers to the Radio Frequency

Identification (RFID) reader’s antenna. The positioning analysis done is on the place-

ment of RFID readers’ antennas to minimise readers’ interference, to offer more cover-

age while conforming to local regulations. The analysis includes but is not limited to

the investigation of in-door path loss models, the computation of signal strength, and

the examination of Specific Absorption Rate (SAR).

“Dual-Frequency Antenna Design of High Frequency Ratio for Advanced Electronic

Code Responding Labels” involves the design of an RFID tag antenna and is not re-

lated to the RFID reader antenna. Frequency ratio is computed through the division

of the higher frequency by the lower frequency and hence is always greater than 1. A

dual-frequency antenna with low frequency ratio (<5) is common, but not for a dual-

frequency antenna with high frequency ratio. The two frequencies of interest in this

thesis are the HF band (13.56 MHz) and the UHF band (860 - 960 MHz) and hence the

intended dual-frequency antenna has a frequency ratio in excess of 70. The work pre-

sented in this thesis includes the design of individual HF tag antennas and UHF tag

antennas, as will be shown in detail in Section 1.5.

1.2 Background of Research

Radio Frequency Identification (RFID) forms the background of this research. An RFID

system, is in fact a very complex system, which involves various fields of study, such as

antenna design, signal processing, integrated circuit design, RF hardware design and

information networking. The focus will be on an RFID system for the supply chain

Page 2

Chapter 1 Thesis Introduction

to track individual items, and thus focuses on either HF or UHF RFID systems using

passive RFID tags.

A simple RFID system is as shown in Fig. 1.1. An RFID tag, which contains a unique

Electronic Product Code (EPC) [4] is attached to an item. Since the tag is passive, an

RFID reader antenna is required to power up the tag. Once the RFID reader detects the

tag, the EPC of the tag is sent to a host computer for processing and business decision

making. For example, if product A is checked out at a sales point, the host computer

will receive the EPC of Product A and when the stock of Product A is running low, the

host computer can automatically generate a purchase order of Product A. A detailed

review of RFID systems is presented in Chapter 2.

RFID Network(EPC Network)

RFIDReader

Antenna RFIDTag

HostComputer

1 2

3

Figure 1.1. A simple RFID system. (1) Host computer to link RFID readers to an RFID Network

(generally an Electronic Product Code (EPC) Network). (2) An RFID reader with one or

more reader antennas connected. (3) Passive RFID tags powered up by reader antennas

so that these tags can be read by the readers.

The research areas involved, according to the societies of the Institute of Electrical and

Electronics Engineers (IEEE), are [5]:

• Antennas and Propagation

• Electromagnetic Compatibility

• Microwave Theory and Techniques

1.3 Motivation

RFID is not a new technology. As will be discussed in Chapter 2, the history of RFID

can be traced back to the early 20th century [6]. However, it is only in the late 1990’s

Page 3

1.3 Motivation

that RFID gained widespread attention when its potential in automating the supply

chain was investigated. It is believed that RFID can revolutionise the supply chain

using real time monitoring, which can offer instant restocking capabilities, elimination

of the overstocking problem, reduction in counterfeiting and many others.

To fully embrace the potential of an RFID system in a supply chain, it is necessary

to attach an RFID tag to every box, carton or pallet in the supply chain. Also, every

RFID tag contains a chip which holds a unique identification code of the object the

tag is attached to. Hence only passive RFID tags are feasible to be deployed in this

manner, or else the total cost of any product with an RFID tag attached will increase to

an unacceptable level.

As mentioned before, the focus of this research is on the deployment of RFID systems

in the supply chain. A supply chain poses a challenging environment for the deploy-

ment of passive RFID tags. Firstly, there are physically thousands to millions of RFID

tags in a relatively small area, such as a warehouse. In other words, a dense RFID

tag environment is common along a supply chain. Intensive research needs to be car-

ried out to ensure good RFID coverage, reliable RFID systems, minimising interference

between tags and readers, while optimising the frequency spectrum in accordance to

local regulations.

Secondly, in a conventional passive RFID system, the reader antenna provides power

to the passive RFID tags before any communication link between a reader and a tag

can be established. The quality of this communication link depends heavily on the

environment. In the worst case scenario, a hostile environment can render an other-

wise functioning RFID system useless. One very interesting point is that ionised liquid

products are very common along the supply chain. Ionised liquid, such as water, can

absorb microwave power easily and hence a UHF RFID system will not perform well

in an ionised liquid filled environment. This prompts an investigation to design a dual-

frequency RFID tag antenna, which can operate at UHF in normal operation and also

when desired at HF when ionised liquid is present.

Page 4


1.4 Contribution

There are two main areas of contribution offered by this thesis with respect to the two

motivations as discussed in Section 1.3.

The first area of contribution is on the dense reader deployment. This thesis presents a

solid and practical way to perform reader antenna positioning analysis. The research

started off with the study of a simple path loss model. This simple model was then

modified to take into consideration environmental factors to cater for different deploy-

ment situations. With a comprehensive model, simulation code was written in MAT-

LAB to predict the effect of RFID readers on each other. This offers an easy way to

examine the interference between readers. Though approximate, as the actual envi-

ronment is dynamic in nature, this allows examination of potential interference before

actual implementation is carried out.

Also, this thesis investigates some novel techniques in synchronising the operation

of RFID readers in a certain vicinity to optimise the performance of an RFID system

when operating under strict local regulations. The local regulation of interest is the

European “Listen Before Talk” provision. The author believes that “Listen Before Talk”

is not beneficial for the widespread adoption of RFID system. Intensive studies carried

out proved that the threshold values as specified in the “Listen Before Talk” provision

are too low to allow wide spread adoption of RFID to be feasible. It is hoped that

the results presented here will strengthen the case for the relaxation of the regulation

imposed upon RFID systems in Europe.

Though not entirely related to those issues discussed above, this thesis extends its

reader positioning study into the implementation of a new technique, named “Sec-

ond Carrier Sensing”. This technique is to prevent the RFID tag from being confused

by multiple interrogation signals from different reader antennas which exist in one

area. Also, included is the examination of Specific Absorption Rate (SAR) in a dense

reader environment. The analysis was carried out using both the path loss model de-

veloped and the worst case scenario, to ensure the deployment of RFID readers would

not create any potential radiation overexposure to anyone.

Page 5

1.4 Contribution

The second area of contribution is concerned with the art of RFID tag antenna de-

sign. This thesis presents the crafting of a novel RFID tag antenna from very basic HF

and UHF tag antennas. The design specifications of the tag antennas in this thesis are

mostly industry driven, where various actual implementations are possible (some were

already implemented in the course of this research). Design decisions are clearly pre-

sented to offer a reference for those interested in tag design for other applications. The

work on individual HF and UHF RFID tag antennas is essential in providing a strong

foundation for the research on a novel dual-frequency antenna, as a dual-frequency

antenna is in fact an antenna with the characteristics of both HF and UHF antennas.

With this foundation, the research on tag antenna design was then extended into the

design of a dual-frequency RFID tag antenna. This dual-frequency antenna supports

both HF band and UHF band operation and has a frequency ratio of more than 70.

This is the first dual-frequency antenna that supports such a high frequency ratio while

maintaining a single feed structure. With this dual-frequency antenna, an RFID tag can

support two very different frequency bands with just a single chip. This is especially

important as HF will offer a better read range in a environment with lots of liquid

products. Indirectly, this will contribute to the effort in spreading and encouraging the

usage of RFID in the supply chain.

Alternative methods in creating a novel dual-frequency antenna of similar character-

istics are also explored, with the advantages and disadvantages of each method being

discussed and compared. This provide a clear picture on the design options of a dual-

frequency antenna for an RFID tag.

Also documented in this thesis are the techniques developed over the the entire re-

search period for measuring the performance of newly designed RFID tag antennas.

This will provide a compact guide for anyone who is interested in the area of RFID tag

design.

A summary of the contribution can be found in the conclusion of this thesis in Sec-

tion 10.4.

Page 6


1.5 Thesis Structure

This thesis is structured into 9 main chapters, excluding conclusion, appendix and

bibliography, and is as shown in Fig. 1.2.

Chapter 2RFID System

Chapter 3RFID Positioning

Analysis

Chapter 5Reader

Synchronisation

Chapter 4Operational

Considerations

Chapter 6RFID Antenna

Design

Chapter 7Dual-frequencyRFID Antenna

Chapter 8Alternative Dual-frequency

RFID Antenna

Chapter 9Antenna

Measurement

Figure 1.2. The complete structure of the thesis.

These 9 chapters can be categorised into 3 major areas:

1. Introduction

• Chapter 1 of this thesis (which is this chapter) explains the title of this thesis

and outlines the background of this research. Also highlighted is the moti-

vation of this research and its contribution to the engineering community.

• Chapter 2 presents a survey of RFID systems and includes a brief history

of RFID, a presentation of a simple RFID system and of variants of RFID

systems. Also discussed are the deployment of RFID in the supply chain,

RFID standards and the future of RFID.

2. Antenna Positioning Analysis

• Chapter 3 is on the RFID reader antenna positioning analysis. Simple path

loss models are discussed and applied in the prediction of the strength of

Page 7

1.5 Thesis Structure

RFID reader interrogating signals. Furthermore, a simulation coded in MAT-

LAB was developed.

• Chapter 4 investigates the operational considerations for RFID deployment

and simulation. It extends the research presented in Chapter 3 and offers

guidelines to avoid pitfalls in setting up simulation software for RFID reader

deployment. Furthermore, the idea of “Second Carrier Sensing” is presented

with its limits discussed in relation to the positions of reader antennas. An

in-depth investigation of Specific Absorption Rate (SAR) is also included,

which is essential before any deployment of dense RFID system.

• Chapter 5 presents the idea of RFID reader synchronisation, where RFID

readers are synchronised to share the limited bandwidth allowed under cer-

tain local regulations. A variation of the synchronisation method is intro-

duced to cater for different needs in real life implementation.

3. Dual-Frequency RFID Tag Antenna Design and Analysis

• Chapter 6 is a large chapter, and mainly deals with the design of RFID tag

antennas. In the beginning of the chapter, the design constraints of small

RFID tag antennas are discussed. The latter parts can be separated into two

major parts, the design of HF RFID tag antennas and the design of UHF

RFID tag antennas.

In the design of HF RFID tag antennas, the theory of a simple HF loop an-

tenna (magnetic dipole) is reviewed. This is followed by the practical design,

simulation, fabrication and testing of HF RFID tags for various applications

which include wine cork tagging and pig ear tagging.

In keeping with above, in the design of UHF RFID tag antennas, the theory

of the simple electric dipole and in particular of its planar form is reviewed,

followed by the practical design, simulation, fabrication and testing of UHF

RFID tags for various applications which include sheep ear tagging and beer

keg tagging.

• Chapter 7 focuses on the design and analysis of a dual-frequency RFID tag

antenna. Though different methodologies are presented, the main focus

Page 8


of this chapter is to design a dual-frequency antenna using the method of

merging an HF antenna and a UHF antenna. Theories for antenna design

were reviewed before actual design, fabrication, and actual tag testing was

carried out. Also included in this chapter, is the development of a minia-

turised version of the designed dual-frequency RFID tag antenna.

• Chapter 8 complements Chapter 7 and explores alternative methods in cre-

ating a dual-frequency antenna, which includes having different feed point

locations or different antenna types when merging an HF together with a

UHF antenna. Furthermore, research is extended into having an HF antenna

acting as a UHF antenna at the UHF band and having a UHF antenna acting

as an HF antenna at the HF band.

• Chapter 9 discusses the measurement techniques used in carrying out all

the measurements to obtain the characteristics of the RFID tag antenna. The

focus was on the technique in obtaining the correct impedance of a UHF tag

antenna. Also included is the analysis on the power transfer between a tag

antenna and the attached chip, when there is mismatch caused by inaccu-

racy in impedance measurement.

1.6 Chapter Structure

A chapter begins with a brief introduction, followed by background study of materials

related to the research and discussion of that chapter. Chapter 2 will be referred to if

any basic RFID background is involved. Most of the important results in any chapter

have been published as conference papers. Any publication used in a chapter will be

mentioned at the end of the introduction section of that chapter or the beginning of a

section.

Page 9

Chapter 2

RFID Systems

THIS chapter presents a study of the fundamental principles of

operation of a passive RFID system. This study and the asso-

ciated literature review provide a strong foundation for the re-

search carried out by the author and will be referred to throughout this

thesis.

Page 11

2.1 Introduction

2.1 Introduction

Radio Frequency Identification (RFID) is a technology which involves several distinct

areas of expertise. As explained in Chapter 1, the focus of research presented in this

thesis is on the RFID antenna positioning analysis in a dense reader deployment zone,

and the antenna design for a dual-frequency antenna. Nonetheless, for completeness, a

detailed background on RFID is presented in this chapter, together with the operational

principles of RFID, that will be referred throughout this thesis.

Firstly, the history of RFID will be presented in the next section, followed by descrip-

tion of a simple RFID system in Section 2.3. Variants of RFID are discussed in Sec-

tion 2.4 though as mentioned in several occasions before, the focus of this thesis in on

passive HF and UHF RFID systems.

The attention is then focussed on the impact of RFID deployment in the supply chain

(Section 2.5) and the current network system responsible in handling the massive amount

of data created by RFID tag reading (Section 2.6). RFID standards and protocols, which

govern the operation of RFID system, are presented in Section 2.7.

Papers on the networking aspect of RFID have been co-authored [7, 8], and the key

points of those papers are duplicated in this chapter.

2.2 History of RFID

RFID, in fact, is not a new technology and has a long history. It is believed that the

birth of RFID can be traced back to the early 20th century, when Radio Detection And

Range (Radar) was invented, where the reflected wave is used to detect an object.

However, it was then discovered that detecting an object is not sufficient in some situ-

ation, especially during the war time, when an identification is also required. Hence,

the first RFID deployment was carried out as a long-range transponder systems of

“Identification, Friend, or Foe” (IFF) for aircraft, so that a friendly aircraft would not

be misidentified as the enemy [9, 6].

Page 12

Chapter 2 RFID Systems

The first patent on RFID is by Harris in 1960, titled “Radio transmission systems with

modulatable passive responder”. In this patent, Harris presented several ways of com-

munication between a reader and a passive responder (tag). [10].

Early journal publications on RFID include:

• “Small resonant scatters in field measurement” by Harrington [11] in 1962, where

a general formulation for a backscattered field from loaded objects is given.

• “Electromagnetic scattering by antennas” by Harrington [12] in 1963, where he

relates the theory of loaded scatterers to antenna theory.

• “Electronic surveillance system” by Cole [13] in 1972, where surface acoustic

waves (SAW) are utilised for the generation within a passive tag of a coded reply,

with electromagnetic communication between the tag and a reader antenna.

• “Identification using modulated RF backscatter” by Koelle, Depp and Freyman

[14] in 1975, where a successful demonstration of identification using RFID backscat-

tering is shown.

However, because of cost and size issues, RFID had limited usage and coverage in the

70’s and 80’s. It was deployed in certain toll collection and object tracking applications.

With the advances in semiconductor technology and miniaturisation in 90’s, the cost

and size of the passive transponders were reduced to a very acceptable level and at the

end of 20th century, they were beginning to be used in the supply chain.

2.3 A Simple RFID System

RFID is a technique used to identify objects by means of electromagnetic waves. An

object can be tagged with an electronic code responding label. An electronic tag con-

sists of an antenna and an integrated circuit. Upon receiving any valid interrogating

signal from any interrogating source, such as a reader, the tag will respond according

to its designed protocol. The relationship between a tag and a reader is illustrated as

Figure 2.1.

Page 13

2.4 Variants of RFID

Reader

Tag

Tag

Tag

TagAntenna

Forward LinkReturn Link

Figure 2.1. RFID system: reader and tag relationship. One or more antennas can be connected

to a reader, and are used to communicate with tags within the field of interrogation.

Signals sent from reader through an antenna to a tag are called forward link signals

while the signals from a tag back to reader are called return link signals.

A detailed description of an RFID system can also be found in RFID Handbook by

Finkenzeller [15].


2.4.1 Active and Passive RFID Tags

There are two types of tags, active and passive. Both active and passive RFID tags are

defined in [16].

An active tag is defined as an “RFID device having the ability of producing a radio

signal”. Normally an active tag has its own battery source; it has a greater read range

compared to a passive tag but is limited by the life time of its battery.

A passive tag is defined as an “RFID device which reflects and modulates a carrier

signal received from an interrogator”. A passive tag is normally energised by an in-

terrogating signal from a reader antenna and does not have any other internal power

source. It typically has shorter read range as compared with an active tag.

There are several good RFID passive tag chip designs available in the literature. For

example, a low power high performance UHF passive tag chip was presented by

Karthaus [17], which requires only 16.7 µW to power up the tag chip and has a read

range of around 10 m. There is also a low cost UHF passive tag design offered by

Page 14


Glidden [18]. Tag chips are mentioned here because knowledge of the tag chip input

impedance is essential in antenna design.

2.4.2 Backscattering

When an electromagnetic wave is radiated in a direction, part of it will be reflected

back to its source if there exists a reflective (metallic) object in its path of propagation.

This is the concept of radar system, where a foreign object in the airspace (such as an

airplane) can detected.

Backscattering is based on the similar concept. The only difference is that in backscat-

tering, the reflected wave is used to transmit data. By changing the impedance of the

reflective object, the reflected wave will be changed accordingly.

In the case of RFID operation, a passive RFID tag consists of two important parts,

the tag antenna and the tag chip. When this tag enters an interrogation zone of an

RFID reader antenna, it will be powered up by the reader antenna. When instructed to

backscatter its unique EPC number, the tag chip will change its input impedance in an

order representing its EPC in binary form (high impedance for “1” and low impedance

for “0”, or vice versa) while the reader is transmitting a continuous wave. The reflected

wave is monitored by the reader and then interpreted into the EPC number of the tag.

Zinput = A

Zinput = A’

Tag

Tag

F

F

R

R’

Figure 2.2. The principle of backscattering. The reflected wave changes depending on the input

impedance of the tag, and can be used to represent binary “1” and “0” to transmit

data. In the figure above, F is the forward wave, while R and R’ are distinguishable

reflected waves depending on the input impedance of the tag.

Page 15


2.4.3 Common Frequency Bands for RFID Operations

There are 4 popular bands in RFID, which are the LF band (less than 135 kHz), the

HF band (13.553 - 13.567 MHz and commonly known as the 13.56 MHz band), the

UHF band (860 - 960 MHz) and the so called microwave band (2.400 - 2.4835 GHz and

commonly known as the 2.45 GHz band) [19]. Each band has positive and negative

characteristics.

The Auto-ID Center has published protocols for HF and UHF RFID operations. LF

has been left out, as under the vision of Auto-ID Center, each tag has to bear an EPC

number, which must be at least 64-bits in length [4], and for low frequency operation,

such as in the LF band, very slow reading in a heavily populated tag environment

would result. Also, LF RFID tags require large antenna components and hence are

difficult to implement and are susceptible to electrical noise, which HF can handle

[20]. Microwave tags can offer comparatively very fast reading, but the performance

of microwave tags will suffer, worse than any other bands described above, in the

presence of liquid or metal [19]. It is very difficult to conclude whether HF or UHF

is better for RFID application in supply chains. This is because in any supply chain,

there are lots of different scenarios, and in some HF is better and in the others, UHF

outperforms HF operation.

The strength of UHF lies in the fact that it offers 4 - 5 m average read range (up to

10 m in state of art RFID systems) and is cheaper to produce in large volume [19]. UHF

tags run at a higher frequency and hence have a higher read rate [21]. In other words,

in a UHF system, more data can be transferred within a fixed period of time, and

hence UHF is often deployed in a situation where more on board data is required [22].

However UHF RFID operation has difficulties in environments that are packed with

liquid or contain lots of metal, or even will not work properly in livestock industries

such as in the area of feedlots or slaughterhouses [23]. This is caused by the nature of

the radio frequency spectrum absorption by water as shown by [24], in which UHF is

highly absorbable.

On the other hand in item level tagging, HF has shorter read range (around 1 m) and

has a slower tag discrimination speed as compared to UHF [22], though it performance

Page 16


is still acceptable with a read rate of more than 200 tags per second. However, HF can

penetrate water better than UHF. According to the radio frequency spectrum absorp-

tion by water shown by [24], the absorption of HF by water is insignificant as com-

pared to UHF. Ref. [21] supports the claim of HF can penetrate water better because of

its longer wave length as compared to UHF, though HF also suffers in the presence of

metal just like UHF. Schurmann also reported that lower frequency HF systems are pri-

marily used in the case where better penetration of objects is needed as quoted by [15].

As a general consensus, both HF and UHF RFID systems are applicable in supply

chains, depending on the RFID application.

2.5 RFID and Supply Chain

As mentioned before, the deployment of RFID systems in supply chains in large num-

bers, with the potential to revolutionise supply chains, has sparked public interest

in RFID technology. A lot of case studies have been carried out by market analysts

[19, 22, 25, 26, 27, 28, 29, 30, 31]. Most of these articles discuss the impact of RFID on

the current supply chain, and how a company can maximise the benefits of an RFID

system.

The formation of the Auto-ID Center in 1999 aimed to develop an open standard ar-

chitecture for creating a seamless global network of physical objects [32]. The ulti-

mate vision was to create an “Internet of things”, and the concepts were well ex-

plained by Cole [25]. With the standardisation of HF protocols (ISO standardisation for

13.56 MHz: ISO/IEC 18000-3 [33] and ISO/IEC 15693 [34]) and UHF protocols (EPC-

global standardisation: UHF Class 1 Generation 2 [35] and EPC tag data standard [4]),

major parties in supply chains were pushing for RFID adoption. For example, one of

the biggest retail stores in the USA, Walmart, mandated the use of RFID in its supply

chains by its top 100 suppliers by Jan. 2005 [36]. Also, the USA Department of Defense

mandated the use of passive RFID in its supply chains by Jan. 2005 [37].

According to a Georgia Institute of Technology business research report, supply chain

problems cost companies between 9 and 20 percent of their value over a six-month

Page 17

2.5 RFID and Supply Chain

period [27]. This shows that the operation of the supply chains cannot afford to be

compromised, and any operational mistake would be costly.

Currently, inventory checking and tracing in a supply chain is done by reading bar-

codes. A barcode, explained in [15], is a binary code comprising a field of bars and

gaps arranged in a parallel configuration. It is read by optical laser scanning, i.e. by

the different reflection of a laser beam from the black bars and white gaps. In other

words, barcode scanning requires manual handling and orientation of each object and

also requires line-of-sight.

A thorough comparison between barcodes and RFID is offered by [15] and supported

by various other independent sources [19, 21, 38, 29]. They claim that, since RFID

does not require line-of-sight, scanning is faster, more effective and hence saves cost.

Furthermore, RFID identifies a unique product item, and enhances Warehouse Man-

agement System (WMS) in dealing with stock replenishment, over-stocking problems

and other supply chain related problems such as counterfeiting.

For example, in an international pharmaceutical supply chain, seven percent of drugs

may be counterfeit. Not only could RFID assure a safer supply of drugs to the con-

sumers, RFID could also save up to 8 billion dollars by 2006 [21]. This is something

that cannot be achieved by barcodes. The USA Food and Drug Administration (FDA)

trusts RFID as a tool to help authenticate drugs, and has taken the initiative in set-

ting up guidelines in using an RFID system in their pharmaceutical supply chain [39].

Also, business cases were presented to support RFID deployment and it is believed

that RFID will offer near perfect inventory information visibility in supply chains [26].

An RFID tag cost as low as 5 cents has been projected [40]. This is also the required cost

of a tag, so that every object in supply chains can be tagged with an RFID tag. Although

this is not the case currently, it is believed that the cost of an RFID tag can be reduced to

that level when the RFID systems are deployed in large numbers and across multiple

applications so that RFID tag production volume is increased to billions of tags.

One of the biggest challenges in RFID large scale deployment is the consumer privacy

issue. Though this is not included in this research, it is mentioned here to show the

author’s belief in RFID policies to avoid any individual privacy invasion.

Page 18


There are many reports on consumers’ fears of RFID deployment in supply chains [22].

Since RFID scanning is a type of automatic scanning [29], or in other words, a person

holding an item will not know when that item is being scanned, consumers may fear

that they are being tracked unlawfully.

The current RFID technology strives to address this issue. The measures taken include

public education on RFID based on the facts that RFID does not have long range scan-

ning capability, and the option that a tag can be “killed” permanently whenever an

item is purchased under endorsed RFID protocols [35].

2.6 The EPC Network

The Electronic Product Code (EPC) Network, originally developed by the Auto-ID

Center with its standards now managed by EPCglobal Inc., was designed and im-

plemented to enable all objects in the world to be linked via the Internet. The EPC

Network has undergone tremendous changes since it was first introduced. At the be-

ginning of this section, the early stage EPC Network will be introduced, followed by

the current state of an EPC Network.

A joint research paper on the EPC Network when it was first introduced had been

published [7, 8], which outlines components of EPC Network and their functionality.

2.6.1 The Early Stage EPC Network

Fig. 2.3 shows the structure of a typical EPC Network in the early stage. The EPC

Network consists of three major components, which are the Savant, the EPC Informa-

tion Service (EPCIS), and the Object Name Service (ONS). Strictly speaking, the reader

should be considered as part of the EPC Network. However, the reader is considered

to be a pure RFID tag interrogator under the control of a Savant, though implementa-

tions of some readers will integrate at least the base functionality of a Savant into the

reader itself.

• Savant

Page 19

2.6 The EPC Network

Figure 2.3. Early stage EPC Network.

The Savant is a middleware system located between a reader (or multiple read-

ers) and the applications in RFID systems. Applications operate on top of, or

within, the Savant operating environment. The Savant passes requests from the

application(s) to the reader(s) and receives unique tag identifiers and possibly

other data from sensors, and passes that information to the application(s). The Sa-

vant has several fundamental functions integrated into its design, some of which

are data filtering, aggregation and counting of tag data. These fundamental func-

tions are required in order to handle the extremely large quantity of data that

RFID systems can generate through the continuous interrogation of tags.

• Object Name Service

The function of the Object Name Service (ONS) in an EPC Network is to iden-

tify the location of the server hosting the appropriate information needed by an

application. In other words, the ONS acts like a “reverse phone directory” as

the ONS uses a number (EPC number) to retrieve the location (of data) from its

database. To encourage rapid development of the ONS, the ONS is purely based

on existing Internet technology and infrastructure. The first generation of ONS

system designs were based upon DNS systems with the first implementations

utilising existing DNS implementations with customised configurations.

• Physical Markup Language

Page 20


The Physical Markup Language (PML) defines the way information is transferred

in the EPC Network system. PML Core is based on the existing standard, XML

Schema Language. It uses “tags” to format the data before the data is sent. PML

does not use short “tags” for data formatting. Although more bandwidth will

be required for data transfer, compared to using short “tags”, the use of longer

descriptive tags increases human readability and hopefully will avoid mistakes

in interpreting and understanding the data, and how that data is to be handled.

• EPC Information Service

EPC Information Service (EPCIS) is the gateway between any requester of in-

formation and the database. It receives and sends messages in PML with any

requester of information, although it’s communication with the database can be

in any format or standard.

2.6.2 The EPC Network Current State

The new EPC Network is relatively different from the original version; it is more de-

fined and better structured. The explanation of the EPC Network at the current state in

this section is based on the EPCglobal Architecture Framework published in 2007 [41].

A whole picture of an EPC Network is as shown in Fig. 2.4.

Basically, there are two distinct types of components in the new EPC Network: soft-

ware/hardware role, and interface. They are self explanatory. Software/hardware role

components are either a piece of software or hardware which handle tasks allocated to

it. Interfaces govern how software/hardware role components communicate to each

other.

All the role components and interface components shown in Fig. 2.4 are explained

in detail in [41]. Some important parts are briefly highlighted to compare with the

original version of an EPC Network:

• Application Level Events

Page 21

2.6 The EPC Network

NOTE: This figure is included on page 22 of the print copy of the

thesis held in the University of Adelaide Library.

Figure 2.4. The Current EPC Network. Reproduced with permission from

[41] ©2006 EPCglobal.

Page 22


Ref. [42] is the complete specification of Application Level Events (ALE). It is an

interface component which handle the data flow of the filtering and collection on

the inventory of RFID tags.

• EPC Information Services

Ref. [43] is the complete specification of EPC Information Services (EPCIS). There

are several interface and role components bearing the name of EPCIS in Fig. 2.4.

Basically any component related to the EPCIS contributes to the data accessing,

sharing and storing between all the authorised RFID networks. By linking RFID

networks together, a global size network with real-time data sharing, also known

as an EPC Network, is made possible.

• Object Name Service

Ref. [44] is the complete specification of Object Name Service (ONS) as an inter-

face component. An ONS interface is used to locate the reference to the EPCIS

service or any other related other services. In Fig. 2.4, there are the Local ONS

and the ONS Root. They are related to but not the same as an ONS interface.

• Reader Management

Ref. [45] is the complete specification of Reader Management. As suggested by

its name, Reader Management is a role component which manages the operation

of RFID readers, mainly on the monitoring of the operating status and health of

RFID readers.

• Reader Protocol

Ref. [46] is the complete specification of Reader Protocol. Reader protocol is an

interface component. Basically, reader protocol allows data from RFID readers to

be passed up the EPCglobal Network stack, for example the ALE.

The basic idea of both the EPC Networks (early stage and current stage) is the same. It

is to enable real-time global data sharing for product tracking. The current EPC Net-

work is an updated version of the early stage EPC Network, retaining all the operating

Page 23

2.7 RFID Standards

principles while offering better clarity. The EPC Network is still evolving, depending

on the needs of RFID adopters.

2.7 RFID Standards

To enable and encourage wide spread adoption of RFID systems, one of the most im-

portant factors is to have RFID systems standardisation. Not only does this simplify

the hardware and software design for RFID systems, it also increases user confidence

in RFID.

Basically, there are two different areas of standardisation. The first one is to regulate the

spectrum usage and is always governed by local regulatory authorities. The common

standards are:

• “Title 47 part 15.247” published by Federal Communications Commission (FCC),

in 2001 [47]. This regulations is mainly adopted by the United States of America.

• “EN 302 208-1 v1.1.1: electromagnetic compatibility and radio spectrum matters

(ERM); radio frequency identification equipment operating in the band 865 MHz

to 868 MHz with power levels up to 2 W” published by European Telecommu-

nications Standards Institute (ETSI) for European Standard (Telecommunications

Series) in 2004 [48]. This regulation is mainly adopted by the countries in the

European Union.

• “AS/NZS 4771:2000 technical characteristics and test conditions for data trans-

mission equipment operating in the 900 MHz, 2.4 GHz and 5.8 GHz bands and

using spread spectrum modulation techniques, incorporating amendment No. 1”

published by Standards Australia.

The second type of standardisation is to have a uniform air interface and command

sets between an RFID reader and an RFID tag so that an RFID reader produced by a

company can be integrated into an RFID network setup by another company. There

Page 24


are two bodies responsible in providing standards, ISO (International Organization of

Standardization) [49] and EPCglobal [50].

The following standards are some of the standards available from ISO:

• For LF RFID operation: ISO 11784:1996: Radio-frequency identification of ani-

mals - code structure [51] and ISO 11785:1996: Radio-frequency identification of

animals - technical concept [52].

• For HF RFID operation: ISO/IEC 18000-3:2004: information technology - radio

frequency identification for item management - Part 3: parameters for air inter-

face communications at 13.56 MHz [33].

• For UHF RFID operation: ISO/IEC 18000-6:2004: information technology - radio

frequency identification for item management - Part 6: parameters for air inter-

face communications at 860 MHz to 960 MHz [53].

• For RFID terminology: ISO/IEC 19762-3, information technology - automatic

identification and data capture (AIDC) techniques - harmonized vocabulary -

part 3: radio frequency identification (RFID) [16].

The following standards are some of the standards available from EPCglobal:

• For the RFID operation in HF band, the standard is documented in “13.56 MHz

ISM band class 1 radio frequency identification tag interface specification: can-

didate recommendation, version 1.1.0” [54]. The HF Generation 2 Tag Proto-

col Standard is currently in development by the HAG HF Air Interface Working

Group.

• For the RFID operation in UHF band, the standard is documented in “EPC radio-

frequency identity protocols class-1 generation-2 UHF RFID protocol for com-

munications at 860 MHz - 960 MHz version 1.1.0” [35]. This version is consistent

with ISO/IEC 18000-6 Type C [53].

Page 25

2.7 RFID Standards

• As every tag contains a unique number and this unique number should be re-

trievable by any reader which conforms the the above stated standards, the tag

data contained by any tag must be standardised too. The standards is “EPC tag

data standards version 1.1 rev. 1.31” [4].

Page 26

Chapter 3

Path Loss and PositionAnalysis

Asimple path loss model for RFID systems is used to analyse

two collision problems in a dense reader environment. The

first problem is the reader interference with tag replies prob-

lem. The second problem is the effect of the “Listen Before Talk” (LBT)

regulations in European countries. It is shown that these two problems will

impede the wide-scale deployment of RFID systems. Based on a suggested

path loss model, a simulation script is written in MATLAB to visualise the

impact of reader interference with tag replies problem (the first problem)

and the effect of LBT regulations (the second problem) when multiple read-

ers are deployed in a same vicinity. With the help of this simulation soft-

ware, antenna positioning analysis was carried out efficiently.

Page 27

3.1 Introduction

3.1 Introduction

In a dense reader environment, there will be multiple RFID readers, and various types

of collision problems will reduce the reliability and efficiency of the RFID system.

RFID readers can use different channels to minimise collision. However, with limited

channels, collision will still occur in dense reader environments. The results can be

unsatisfactory with respect to read times and an unacceptable level of misreads may

result. In the worst case, collision problems could paralyse an entire RFID system.

Furthermore, collision problems also involve the disruption of the operation of other

RF devices in nearby frequency bands. Regulatory bodies have set some strict restric-

tions on RFID radiation to minimise this disruption problem. The concept of “Listen

Before Talk” (LBT) has been included in the European regulation, ETSI EN 302 208 [48],

which places severe restrictions on RFID deployment and causes some uncertainties

over the feasibility of RFID global deployment.

There are many types of collision problem and the following list shows some of the

more common types:

1. Reader Interference with Tag Replies (RITR), where transmissions from one reader

antenna interfere with tag reply signals in the receiver antenna of another reader.

2. Reader Induced Tag Confusion (RITC), where two reader antennas send simul-

taneous interrogating signals to a tag.

3. Regulatory Reader Shutdown (RRS), where a reader antenna is not allowed by

local regulations to transmit in a channel when there is a signal with a certain

strength detected in that channel.

The short-form names of these common collision problem, i.e. RITR, RITC or RRS, will

be used in this thesis. An alternate description of RFID collision problems can also be

found in [55].

There are two major objectives in this research:

Page 28

Chapter 3 Path Loss and Position Analysis

• To report on the analysis of the Reader Interference with Tag Replies (RITR) prob-

lem and provide a solution to RFID deployment regarding the RITR problem for

the benefit of those eager to set up RFID systems.

• To provide an analysis on the Regulatory Reader Shutdown (RRS) problem. The

focus will be on the restrictions imposed by “Listen Before Talk” to the placement

of RFID reader antenna.

The research performed to achieve these objectives is explained in the following se-

quence of elements:

1. A comprehensive background study on path loss models, with a decision made

later to use a piece-wise linear model for path loss in the RFID context:

Section 3.2 explains the theory of a path loss model, and shows how an in-building

path loss model is related to a free space model.

Section 3.3 explains the simple path loss experiments carried out around the lab-

oratory, and includes discussion of results.

Section 3.4 forms a path loss model, based on the results in Section 3.3, for the

RFID context.

2. A simple study of the Reader Interference with Tag Replies (RITR) problem based

on the chosen path loss model:

The path loss model chosen in Section 3.4 is used to investigate a simple two-

reader RITR situation in Section 3.5. Also, the safe distance between two RFID

reader antennas, before RITR occurs in a same channel, is computed and dis-

cussed.

3. Development of a simulation software based on the chosen path loss model:

Section 3.6 discusses the theory of power density, antenna gain pattern of an RFID

antenna, frequency channelling and the concept of “Listen Before Talk”. This

section is mostly informative but provides a strong background study for the

RFID positioning analysis discussion from Section 3.7 to Section 3.9.

Page 29

3.1 Introduction

Section 3.7 explains the concepts behind the simulation software developed in

MATLAB, and also, the input parameters for the simulation software to produce

the desired results.

4. Simulation results and discussions of the reader interference problem:

Section 3.8 re-performs the simple simple two-reader RITR in Section 3.5 using

the developed software from Section 3.7. The strength of this software over the

direct calculation method is examined.

5. Simulation results and discussions of the Regulatory Reader Shutdown (RRS)

problem:

Section 3.9 includes both the simulation results and discussions of various an-

tenna positioning strategies regarding the RRS problem. The focus will be on

the European “Listen Before Talk (LBT) provision. Intensive simulations were

carried out to visualise RFID reader signal strengths as received by isotropic or

directional receivers at various distances with respect to transmitting antennas,

either operating in the same channel or in various neighbouring channels.

6. Examples of some real life applications:

Some real life applications including the potential of reader synchronisation in

handling problems that may arise in a dense reader environment are presented

in Section 3.10.

It is hoped that this research can provide sufficient and useful guidelines on the safe

distance between antennas in a dense reader environment (to resolve the RITR prob-

lem) or even in a “Listen Before Talk” regulated area (to resolve the RRS problem). It

is also the belief of the author that by careful planning of antenna positioning, an RFID

system can be optimised and its interference to other RF systems in the surrounding

area can be minimised.

This chapter is the merging of two published papers [56, 57], with the study of reader

interference problem coming from [56] and the work on LBT analysis coming from [57].

Page 30


3.2 Path Loss Model

3.2.1 Free Space Path Loss

For a pair of lossless antennas in free space with optimum orientation we may write

the power transfer ratio in the form:

Pr

Pt= gtgr

(λ

4πd

)2

(3.1)

where λ = wavelength; Pt = transmitted power; Pr = available received power; gt =

transmitter antenna gain; gr = receiver antenna gain; d = separation distance between

antennas.

For some purposes it is desirable to separate the effects of antenna gain and distance

between antennas, and give the name free space path loss to the remaining factor in

the above equation. By expressing this factor in dB we have the free space path loss

expression:

PL(dB) = 20 log10

(4πd

λ

)(3.2)

Using the frequency of f = 915 MHz, λ = cf = 0.33 m, for a separation d = 1 m, PL(dB)

= 31.61 dB.

Also, (3.1) can be reexpressed as:

Pr(dBW) = Pt(dBW) + gt(dBi) + gr(dBi) − PL(dB) (3.3)

where Pt = transmitted power; Pr = available received power; gt = transmitter antenna

gain; gr = receiver antenna gain; PL = path loss.

3.2.2 In-building Path Loss

An in-building path loss is a path loss that occurs in a physical building, and it is

not the same as a free space path loss, as an in-building path loss will normally take

Page 31

3.3 Experiment

into consideration path obstruction, reflection, absorption and other attenuation effects

introduced by the presence of objects inside a building. Free space path loss has a

simple d−2 variation but in-building path loss has a more complex structure. The in-

building path loss model chosen for the purpose of simulation is given in (3.4), and is

explained by Rappaport [58]:

PL(dB) = PL(d0) + 10 × n × log10

(dd0

)(3.4)

where d0 = an arbitrary reference distance; n = a value that depends on the surround-

ings and building types; and d = the separation distance between the two antennas.

d0 will be chosen as 1 m and PL(d0), which is the in-building path loss at 1 m away, will

be approximately the same as the free space path loss at 1 m away, as room reflections

are not huge at this small distance. Hence, from (3.2), PL(1 m) is approximately 32 dB,

and (3.4) becomes:

PL(dB) = 32 + 10 × n × log10

(dd0

)(3.5)

As explained in (3.4), n is a empirical value that depends on the surroundings and

building types, and is only obtainable through experiment. An environment with high

n is a hostile environment for radiation, and its in-building path loss will be higher

when compared to the case of a low n environment.

3.3 Experiment

An interrogating RFID antenna was set to transmit a known signal while a measuring

spectrum analyser connected to a receiver antenna was moved away from the trans-

mitting antenna. The strength of the received signal was recorded versus the distance

away from the transmitting antenna. Removing antenna gain from the measured val-

ues gives us the values of path loss. The transmitting and receiving antenna used in

this experiment both have a gain of 6 dBi. Fig. 3.1 is plotted using logarithmic scale,

and we have a straight line approximation of:

Page 32


PL(dB) = 31 + 40 log10

(d1

)(3.6)

for d > 8 m. The results resemble the model based on (3.7) in Section 3.4 and also

strengthen the belief that reader interference problem must be solved for large scale

RFID deployment.

0

20

40

60

80

100

120

1 10 100

Distance of Separation Between Antennas, d (m)

Pat

h L

oss

(dB

)

Figure 3.1. Experimental path loss results. Computed path loss results from experiment, with

variation in the distance between the transmitting and receiving antennas.

The in-depth details of the experiments, including equipment setup, experiment venues

and data analysis can be found in Section 4.4 and Appendix F.

3.4 Path Loss Model for RFID

For the theoretical calculations in this chapter, we chose n = n2 = 3.5, when the separa-

tion is more than or equal to 8 m, and we chose n = n1 = 2.5 when the distance is less

than 8 m. i.e. we are using approximation equations:

PL(dB) =

⎧⎪⎪⎪⎨⎪⎪⎪⎩

PL(d0) + 10 × n1 × log10dd0

0 < d ≤ 8 m

C + 10 × n2 × log10dd0

d ≥ 8 m

(3.7)

Page 33

3.5 Simple Two Reader Interference

where PL(d0) is 31.61 dB at 915 MHz (from (3.2)). For all the analysis presented in this

thesis, PL(d0) takes the approximate value of 32 dB. C is constant value with respect

to the chosen PL(d0), n1 and n2 values to ensure continuity of (3.7) at position d = 8 m.

For a PL(d0) of 32 dB, n1 of 2.5 and n2 of 3.5, C takes the value of 22.97 dB.

Hence, effectively, the in-building path loss model from (3.5) had been modified to the

piece-wise linear (in log10dd0

) relation given in (3.7) and plotted as shown by the heavy

line in Fig. 3.2.

0

20

40

60

80

100

120

140

1 10 100

Transmitter-Receiver Separation (m)

Pa

th L

oss (

dB

)

n=4

n=3

n=2

n=6n=5

Figure 3.2. Plot of proposed piece-wise linear in-building path loss model against distance.

Dotted lines represent the path loss model with environment factor, n, from 2 to 6

respectively, calculated using (3.5). The bold line represents the suggested piece-wise

linear model, represented by (3.7).


As highlighted in the introduction section of this chapter, one type of reader collision

chosen for study is the Reader Interference with Tag Replies (RITR) problem. RITR

happens when transmitting signals from a reader antenna drown reply signals of a tag

to another reader antenna.

In this section, a scaled down version of RITR is discussed, where the RITR involves

only two readers, and the readers are assumed to transmit and receive in the same

Page 34


channel of a multi-channel frequency band. Consider the case where there are two

readers, A and B, using the same channel, Channel C. It is assumed that Reader A and

Reader B are identical, the antennas for both of them are the same and have the same

gain. Also, both of the antennas are facing each other. Reader A uses Channel C to

interrogate a tag (which is located directly in front of antenna of Reader A), and the

tag will have in-band backscattering in response to Reader A. If the power delivered

to both reader antennas is 0 dBW, with reader antenna gains of 6 dBi, a total of 6 dBW

EIRP will be obtained. Fig. 3.3 shows diagrammatically how the interrogation between

Reader A and the tag occurs. Signal paths are shown using the bold dotted lines. At

1 m away, the path loss for signals 1 and 2 obtained using (3.7) is 32 dB. Hence the total

path loss is approximately 64 dB (2 × 32 dB).

Reader A Reader BTag

3

1

2

Reader

Antenna

Figure 3.3. A simple illustration of Reader Interference with Tag Replies (RITR). RITR will

occur when a reply from a tag is interfered with by a signal sent from another nearby

reader. In the example illustrated, tag reply 2 is interfered with by interrogation signal

3 from Reader B. The bold dotted line (signal path) signifies that the two reader

antennas are facing each other, while the tag is in-line with the maximum gain direction

of antenna A. The numbered arrows indicate the distances and directions of signal

propagation between objects of interest. The numbered arrows do not represent the

actual propagation paths.

The tag antenna has a gain of approximately 1.5. However, to take in consideration

a probable tag polarisation misalignment, i.e. the tag antenna is not oriented in the

most optimal way, it is assumed that the tag has a unity antenna gain. Also assumed,

is that the efficiency of the tag is 10%, and the signal will suffer another 10 dB loss.

All the losses (path loss + tag efficiency-tag antenna gain) are summed up to be 74 dB.

Since Reader B is also using the same channel, Channel C, the interrogation signal sent

by Reader B will interfere with the in-channel backscattered signal from the tag. The

Page 35


Table 3.1. The effect of tag distance on multi-reader interference. Referring to Fig. 3.3,

Dtag−A is the distance between the tag and reader antenna A; MinB int is the minimum

distance between the reader antenna A and reader antenna B, before the signal from

reader antenna B interferes with the tag reply to antenna A; Df s is MinB int in free

space.

Dtag−A (m) Total Loss (dB) MinB int (m) Df s (m)

1 74 28.7 130.7

2 89 76.9 735.2

5 109 286.5 7,352.5

10 123 718.3 35,237.1

question is how near Reader B needs to be to interfere with the backscattered signal.

The comparison is made between (a) path loss of path 1 and path 2, tag antenna gain,

and tag efficiency loss, and (b) path loss of path 3. This comparison is only applicable

in the situation as described in this section. In this section, the losses in both (a) and (b)

are called Total Loss. Of course in the usual scenario, the values are different for the

two cases.

Note that the gains of reader antennas (for both Reader A and B) should be considered

as well. However, since they contribute a total of 12 dB to both case (a) and case (b),

they are not included in the comparison calculations.

Path 3 in Fig. 3.3 shows the distance and direction of signal travelling along the dotted

line from Reader B to Reader A. Using (3.7) to calculate path loss, a distance of around

28.7 m is needed to have path loss, or total loss in this case, of 74 dB. Table 3.1 shows

some results on the minimum distance for Reader B to interfere with the tag reply.

Again, the results in columns 2 and 3 of Table 3.1 are computed using (3.7), and (3.7)

takes into consideration in-building propagation loss. It is very natural to also raise

the question of what will be the case, if free space propagation loss is considered. The

results obtained using free space propagation loss model are attached as column 4 in

Table 3.1.

Page 36


It is discovered that, if a tag is located 10 m away from an interrogating reader, the

antenna of the interfering reader must be around 718.3 m away. Since a state of art

reader in the market can have a read range of around 10 m when reading a passive tag,

to put the next reader several hundred metres away is not sensible. Section 3.10 in the

latter part of this chapter provides ideas to solve this problem.

We continue to make the assumption that free space path loss is not applicable in prac-

tical applications, and that at the distances emerging from these calculations, at least

some obstacles will be present, and a within-building propagation loss model is ap-

propriate. As shown in Table 3.1, when free space path loss model is used, the safe

distance between two antennas facing each other to prevent interference becomes rel-

atively huge (> 35 km), but we do not believe this model is relevant to practical situa-

tions.

3.6 Background for RFID Positioning Analysis

This section presents important background studies which assist the detailed discus-

sion of RFID positioning analysis. Firstly the difference between EIRP (Equivalent

Isotropic Radiated Power) and ERP (Effective Radiated Power) is discussed, followed

by the examination of a frequency spectrum channelling technique used in RFID oper-

ation. Also included is the in-depth investigation of the “Listen Before Talk” provision

in the European RFID regulation.

3.6.1 Power Density

Gain of an antenna is the ratio of the radiation intensity, in a given direction, to the

radiation intensity that would be obtained if the power accepted by the antenna were

radiated isotropically.

Power density, S, which at a distance d for a transmitting antenna in the direction of

maximum gain, has the value:

Page 37


S =(

EIRP4πd2

)(3.8)

where Equivalent Isotropic Radiated Power (EIRP) is defined as, in a given direction,

the gain of a transmitting antenna multiplied by the net power accepted by the antenna

from the connected transmitter. In simple term, EIRP is the product of power delivered

to the antenna (Pt) and antenna gain with respect to the gain of an isotropic radiator

(gt):

EIRP = gtPt (3.9)

Effective Radiated Power (ERP) is a more popular term in Europe and is defined as,

in a given direction, the relative gain of a transmitting antenna with respect to the

maximum directivity of a half-wave length dipole (or alternatively the maximum gain

of a lossless half-wave length dipole) multiplied by the net power accepted by the

antenna from the connected transmitter. In simple term, ERP is the product of power

delivered to the antenna (Pt) and antenna gain with respect to the maximum gain of a

lossless half-wave length dipole (gd):

ERP = gdPt (3.10)

It should be noted that the “E” in EIRP stands for equivalent while the “E” in ERP

stands for effective. Since the maximum gain of a lossless half wave length dipole

relative to an isotropic radiator is 1.64, gd can be expressed in term of gt:

gd =gt

1.64(3.11)

With (3.11), ERP can be expressed in term of gt:

ERP =gtPt

1.64(3.12)

Thus, by dividing (3.9) and (3.12), and rearranging:

Page 38


ERP = 1.64 × EIRP (3.13)

Hence, (3.8) can be reexpressed as:

S =(

1.64 × ERP4πd2

)(3.14)

3.6.2 Antenna Gain Pattern

As shown by (3.1), antenna gain must be taken into consideration, as well as path loss,

to relate the power delivered to a transmitting antenna to power received by another

antenna. A typical RFID antenna is a directional circularly polarised antenna with a

gain of 6 dBi. The polar plot of the gain of a typical RFID antenna is shown in Fig. 3.4,

and has been used, when directional antennas are called for, to simulate and analyse

the antenna positioning in the dense reader environment presented in this thesis.

30

210

60

240

90

270

120

300

150

330

180 0

6 dBi 10

0

−10

−20

Figure 3.4. Polar plot of the antenna gain of a directional circularly polarised RFID antenna

with a gain of 6 dBi. This plot is not a direct measurement plot. An anechoic chamber

measurement was carried out on a 6 dBi RFID antenna. The measurement results are

then sampled at discrete points before input into MATLAB for plotting.

Page 39


3.6.3 Frequency Spectrum Channelling

Frequency spectrum channelling is a popular technique used in reducing both types

of reader collision problem considered in this chapter. One of the biggest challenges

in frequency spectrum channelling is the sideband interference, where the sideband

of a particular transmitting channel interferes with signals in a neighbouring channel.

Under the latest EPCglobal UHF Class 1 Generation 2 protocol [35], the sidebands of

a transmission must adhere to the limits as shown in Fig. 3.5. The dBch is defined as

decibels referenced to the integrated power in the reference channel. The simulation

software developed also uses this limit to investigate interactions of readers operating

using different frequency channels. This investigation is presented in detail in Sec-

tion 3.9.4.

There are two different types of transmit mask, one for the so-called multiple-interrogator

environment and the other for the so-called dense-interrogator environment. Accord-

ing to [35], the multiple-interrogator environment is defined as an operating environ-

ment (in free space, a sphere with radius of approximately 1000 m) within which a

modest number of the available channels are occupied by active interrogators (for ex-

ample, 5 active interrogators operating in 25 available channels). The dense-interrogator

environment is defined as as an operating environment (in free space, a sphere with

radius of approximately 1000 m) within which most or all of the available channels are

occupied by active interrogators (for example, 25 active interrogators operating in 25

available channels).

Note that in explaining the difference between a multiple-interrogator environment

and a dense-interrogator environment, [35] defines a sphere with radius of approxi-

mately 1000 m in free space as the operating environment. Referring to (3.7), the path

loss at 1000 m away using n1=n2=2 (free space) is approximately the same as the path

loss at 100 m away when n1=2.5 and n2=3.5 (the suggested path loss model for RFID).

Hence the operating environment can be restated as a sphere with radius of approxi-

mately 100 m in real life.

As stated before, more details on frequency spectrum channelling studies will be pro-

vided in Section 3.9.4.

Page 40




Figure 3.5. Transmit mask for multiple-interrogator and dense-interrogator

environments.

Reproduced with permission from [35] ©2006 EPCglobal. dBch is defined as the

decibels referenced to the integrated power in the reference channel. Tari is the

reference time interval for a data-0 in Interrogator-to-Tag signalling, and is between

6.25 µs and 25 µs. fc is the centre frequency of a transmission.

3.6.4 Listen Before Talk

ETSI 302 208 is a European regulation governing the operation of RFID readers [48].

It allocates the frequency band of 865 to 868 MHz for RFID deployment. This

frequency band is then divided into fifteen sub-bands or channels, each spanning a

total of 200 kHz. However, when a reader is operating at the maximum radiated

power, which is 2 W ERP (Effective Radiated Power), only ten sub-bands are

available, while the remaining five are utilised as guard bands or for lower power

readers. ETSI 302 208

Page 41


also introduces the concept of “Listen Before Talk”. An extract from the ETSI 302 208

best describes the essence of “Listen Before Talk”. It states “Prior to each transmission,

the receiver in the interrogator shall first monitor in accordance with the defined listen

time for the presence of another signal within its intended sub-band of transmission.

The listen time shall comprise a fixed period of 5 ms plus a random time of 0 ms to

5 ms in 11 steps. If the sub-band is free the random time shall be set to 0 ms” [48]. The

threshold to determine the presence of another signal within the intended sub-band is

shown in Table 3.2. The measurement method is defined in the standard.

Furthermore, once a sub-band has been selected, the RFID reader is permitted to use

that sub-band for up to 4 s. After use, it must free the sub-band for at least 100 ms.

A reader can however, listen to another sub-band for 5 ms and if free use that new

sub-band immediately.

Table 3.2. UHF RFID reader radiated power with corresponding threshold values for LBT,

and minimum allowable distance between antennas. Threshold value obtained from

[48]. Minimum path loss and distance required between a transmitting and a receiving

antennas are calculated with the assumption that both antennas are isotropic radiators.

ERP (W) ERP (dBW) Threshold (dBW) Path Loss (dB) Distance (m)

Up to 0.1 Up to -10 ≤-113 103 193.5

0.1 to 0.5 -10 to -3 ≤-120 117 485.9

0.5 to 2.0 -3 to 3 ≤-126 129 1070.1

Similarly to the results shown in column 2 and 3 of Table 3.1 in Section 3.5, the “Dis-

tance” column in Table 3.2 is computed using (3.7), where we consider an in-building

propagation model with n value set to 2.5 for distance less than 8 m and n value set

to 3.5 for distance more than 8 m. As mentioned before, the value of n changes from

building to building. The path loss model chosen for RFID operation is just a good

approximation.

The main point here is that if RFID readers are going to be deployed on a large scale,

most likely the system will not work in optimal operation mode as the Regulatory

Page 42


Reader Shutdown (RRS) problem will occur. This is due to the fact that LBT will effec-

tively shut down many of the channels, though those channels might have been, in the

absence of the LBT provision, freely available for interrogation between readers and

tags. This RRS problem is one of the collision problem in RFID deployment area which

is covered in this chapter. More details on the study of RRS problem will be provided

in Section 3.8.

3.7 Simulation Concepts

Section 3.5 shows that a path loss model can be used to analyse the Reader Interference

with Tag Replies (RITR) problem. However, as explained in Fig. 3.3, the two reader

antennas are positioned face-to-face and the maximum gain is used in both paths of

the calculation and in consequence has no effect. If the reader antennas are of different

orientations, the calculation will be very tedious and the antenna gain pattern has to

be referred to all the time. Hence, a general purpose simulation software was designed

to handle more complex situations, with advanced features that include:

• Handling of more than two reader antennas

Theoretically, the software can simulate an unlimited number and variety of an-

tennas. However, the computational time will increase linearly with linear incre-

ment of the numbers of reader antennas.

• Full consideration of the antenna gain pattern

The antenna gain pattern for every type of antenna needs to be imported to the

simulation software, but only once. After that, the simulation software can con-

sider the gain pattern for any antenna configuration.

• Results in graphical form

The simulation results will be presented graphically. This makes the interpreta-

tion of the results easy.

The simulation software is used to contribute to the reduction of two distinct collision

problems in dense RFID reader environment:

Page 43


1. The Reader Interference with Tag Replies (RITR) problem.

Similarly to Section 3.5, the simulation software can compute the safe distance

between reader antennas to avoid the RITR problem.

2. The Regulatory Reader Shutdown (RRS) problem in European countries.

As discussed in Section 3.6.4, European countries are subjected to the provision

of “Listen Before Talk” (LBT). The simulation software is used to demonstrate the

constraints in deploying RFID reader antennas introduced by LBT.

The simulation script used is written in MATLAB code to show the power received

by an isotropic antenna (0 dB gain) at any location. The simulation concepts can best

be described with the help of Fig. 3.6. Firstly, a grid with size defined by the user is

formed. The centre of each grid box (not the intersection points of grid lines) can be

used as the location of one or more transmitting reader antennas or as the location of

an isotropic receiving antenna. In normal simulation, only the positions of transmit-

ting antennas need to be specified, while all other unassigned grid boxes will become

isotropic receiving antennas. For example in Fig. 3.6, a transmitting antenna is located

at position (4,1) and is symbolised by × and the rest of grid boxes become (unless

otherwise stated) isotropic receiving antennas, which are symbolised by •.

The resolution setting in the simulation script defines the distance between two neigh-

bouring isotropic receiving antennas (distance between the centres of two horizontally

or vertically neighbouring grid boxes). By default the resolution is set to be 1 m. An-

other common resolution used is 10 m. The resolution can also be set lower than 1 m.

The choice of resolution depends on the accuracy requirement for results and also de-

pends how wide is the solution space. For example, it is not wise to have resolution of

0.01 m and the solution space of 1000 m × 1000 m.

All the transmitting antennas used in the simulation are directional circularly polarised

antennas (unless otherwise stated), with antenna gain pattern shown in Fig. 3.4, and

are located 1 m above ground level (refer to Fig. 3.7 for visualisation of the transmitting

antenna configuration). Hence the grid shown in Fig. 3.6 lies in the x-y plane shown in

Fig. 3.7.

Page 44


Extended

Grid

Extended

Grid

Figure 3.6. Grid used in simulation software. The symbol × represents a possible location of a

reader antenna while the symbols • indicates the positions where the received powers

of an isotropic receiver antenna are simulated.

The software uses the path loss model as shown in (3.7). By specifying the power

delivered to the transmitting antenna, (3.3) is used to compute the received power with

the gain specified in Fig. 3.4, path loss from (3.7), and a unity gain for the receiving

antenna.

The simulation result (example Fig. 3.8) is in graphical form and is able to show the

power received by an isotropic antenna (0 dB gain) at any location, bounded by the

solution space set by the user. A broader solution space means more simulated points

and requires longer simulation time.

The orientation and the position of a transmitting antenna with respect to the simu-

lation grid can be illustrated in Fig. 3.7. The simulation results represent the signal

strength received by an isotropic antenna in the x-y plane.

In summary, the developed software can take in the following input:

Page 45


1 m

Default AntennaConfiguration

Ground

z

yx

Figure 3.7. The orientation and the position of a transmitting antenna with respect to the

simulation grid. The simulation software computes the signal strength at the x-y plane.

By default the x-y plane is 1 m above the ground plane. In Section 3.9.3, the software

is modified so that the elevation of the x-y plane can be changed to any height with

respect to the ground level.

• Position of transmitting antennas: The position of each individual transmitting

antenna is input as cartesian coordinates (x,y). In theory there is no upper limit

on the number of antennas which can be included in the simulation. Nonetheless

the computation time increases in a linear fashion with the increment of the total

number of transmitting antennas.

• Radiation pattern: The radiation pattern of each individual transmitting antenna

is input as a set of discrete values. Unless otherwise stated, for all the simulations

presented in this thesis, the radiation pattern as shown in Fig. 3.4 is used.

• Power delivered to the transmitting antenna: This the power input into the reader

antenna. By default, 1 W is used.

• Solution space: This parameter determines the simulation area. For example, the

area of simulation can be set to 10 m × 10 m.

• Resolution: Resolution setting controls how frequent (with respect to distance)

the simulation software computes the signal strength within the solution space.

For example, the software can calculate every 0.1 m or every 1 m.

Page 46


3.7.1 Simulation Concepts Example

5 10 15 20 25 30 35 40 45 50

5

10

15

20

25

30

35

40

45

50−64.5

−61

−57.5

−54

−50.5

−200

−180

−160

−140

−120

−100

−80

−60

−40

−20

0

Around−52 dBW

Figure 3.8. Simulation results verifying the functionality of the software developed in MAT-

LAB. Results from simulation for simulation concepts example (vertical and horizontal

axes in m, representing the y and x axes shown in Fig. 3.7, received power in dBW). The

simulated antenna is located at position (25,25) facing in the y direction. At a distance

of 10 m away, the signal strength is approximately -52 dBW, which is as predicted by

the path loss model.

Fig. 3.8 shows a simple example of simulation results. In this example, the default

value of 0 dBW of power is fed through a 6 dBi gain antenna of pattern shown in

Fig. 3.4, resulting in a 4 W EIRP radiation. The 6 dBi gain is only valid in the front

direction of the antenna. Hence, at a distance of 10 m away directly in front of the

antenna, the path loss will be 58 dB, as computed from (3.7). The received power at that

location as detected by a 0 dB gain antenna is hence 0 dBW (transmit power) + 6 dBi

(transmitter antenna gain) -58 dB (path loss) + 0 dBi (receiver antenna gain), giving a

predicted received power of -52 dBW. In Fig. 3.8, a horizontal guideline is added to

show the point located 10 m away from the antenna. At that point, the received power

is calculated by the simulation software to be approximately -52 dBW, which agrees

with the more directly calculated result.

Page 47

3.8 Results and Discussion (RITR)


In this section, the results obtained in Section 3.5 (using direct calculation) are repro-

duced using a modified form of the simulation software described in Section 3.7. The

discussion also shows the strength of the developed software over the direct calcula-

tion method. The usage of this same software in predicting Regulatory Reader Shut-

down (RRS) problem is presented in next section. This section only focuses on the

Reader Interference with Tag Replies (RITR) problem.

3.8.1 Modification of Existing Software

As the simulation software described in Section 3.7 is in a very general form to cater

for various applications, a few modifications were carried out to compute the effect of

tag distance on the RITR problem. The normal input parameters (from Section 3.7) are

still required. The additional inputs required are:

• P1(x,y,d): Position and direction of a first transmitting reader antenna (represent-

ing antenna A in Fig. 3.3). d is the direction indicator of the antenna. Currently

the antenna can either point in the x or in the y direction (Refer Fig. 3.7). This is

the antenna experiencing interference problem.

• T(x,y): Position of the tag of interest (representing the tag of interest in Fig. 3.3).

It is assumed that this tag has unity gain in all directions. This tag is sending

replies to the first transmitting reader antenna.

• P2(x,y,d): Position and direction of a second transmitting reader antenna (repre-

senting antenna B in Fig. 3.3) d is the direction indicator of the antenna. Cur-

rently the antenna can either point in the x or in the y direction (Refer Fig. 3.7).

This is the antenna producing interference signals to the first transmitting reader

antenna.

With the above input values, the modified software will execute the following steps:

Page 48


1. Run a simulation with the specified solution space, specified resolution, with the

first transmitting antenna located at position P1(x,y) and pointing direction P1(d),

with default gain pattern (Fig. 3.4).

2. The signal strength as received by an isotropic receiver at position T(x,y) is recorded.

As it is assumed that tag has an efficiency of approximately 10%, this recorded

signal strength is further reduced by 10 dB, and is stored as temptag.

3. Run a second simulation with the specified solution space, specified resolution,

with the first transmitting antenna removed, but with an isotropic radiator lo-

cated at T(x,y) and power delivered to this isotropic radiator set to be temptag.

The signal strength received by an isotropic receiver at position P1(x,y) is recorded

as Stag. The angle between the tag position T(x,y) and the position of the first

transmitting antenna P1(x,y) is then computed to obtain the right gain value for

the first transmitting antenna. This gain is used to increase the Stag value. This

Stag represents the strength of tag replies as received by first transmitting an-

tenna. Note that first transmitting antenna does not appear in this second run of

simulation but the position of this first transmitting antenna, P1(x,y), is required.

4. Run a third simulation with the specified solution space, specified resolution,

with only the second transmitting antenna located at position P2(x,y) and point-

ing direction P2(d), with default gain pattern (Fig. 3.4). The signal strength re-

ceived by an isotropic receiver at position P1(x,y) is recorded as Sreader. The angle

between the second transmitting antenna and the first transmitting antenna is

then computed to obtain the right gain value for the first transmitting antenna.

This gain is used to increase the Sreader value. This Sreader represents the strength

of signals from the second transmitting antenna as received by the first transmit-

ting antenna.

5. Stag and Sreader will be the output values. When Sreader > Stag, there will be a

Reader Interference with Tag Replies (RITR) problem.

It can be seen that antenna gain pattern is referred to frequently in the calculation,

especially when the two transmitting antennas and tag of interest are not aligned in a

Page 49


straight line, as they were in Fig. 3.3. This software by taking into account the antenna

gain pattern and can give a quick indication whether a RITR problem will happen or

not based on the position of the two transmitting antennas and the position of the tag

of interest.

3.8.2 Guidelines to Avoid RITR

In common practice, the relative position and direction of two RFID reader antennas

is in one of the forms shown in Fig. 3.9. Hence it is the aim of this section to provide a

safe distance guideline for antennas in the situations shown in Fig. 3.9 to avoid RITR.

Referring to the antenna gain pattern shown in Fig. 3.4, the gain of the antenna in

front, behind and at the sides are, 6 dBi, -22 dBi, and -14 dBi respectively. The safe

distances for these configurations are obtained through simulations and are presented

in Table 3.3.

The safe distance between two antennas with a angle between them (such as the one

shown in Fig. 3.9 (h), depends heavily on the radiation pattern, the direction of maxi-

mum radiation, and the angle, θ. Hence, the safe distance cannot be presented here as

a guideline but can only be obtained on a case by case basis through simulation.

By following the safe distance guideline tabulated in Table 3.3, RITR can be reduced

but is not avoided. For example, if two 6 dBi directional antennas are to be positioned

back to back (Fig. 3.9 (d)), and tags are to be read not more than 1 m away, the safe

distance between these two antennas is 0.3 m. From the same table, it can be noticed

that it is not advisable to design an RFID system with tags to be read at a far distance

away, as an increased distance between a tag and a reader antenna interrogating the

tag will result in an increment in the safe distance between reader antennas to prevent

the RITR problem.

Page 50


(a) (b) (c)

(e) (f)

(g) (h)

(d)

Figure 3.9. Common positionings of antennas with respect to each other. The coloured

antenna is referred as the first antenna. (a) antennas facing each other; (b) antennas

side-by-side facing in the same direction; (c) antennas side-by-side but facing in opposite

directions; (d) antennas back-to-back; (e) antennas in a line facing in the same direction;

(f) second antenna facing the side of the first antenna; (g) second antenna backing the

side of the first antenna; (h) other positions with an angle between antennas, and

radiation directions as shown.

Page 51

3.9 Results and Discussion (RRS)

Table 3.3. Safe distances, in m, for different antenna positioning to avoid the Reader In-

terference with Tag Replies (RITR) problem. The different antenna positionings are

defined in Fig. 3.9. All these results are obtained through simulations using the simu-

lation script described in Section 3.7. Note that these results are for cases where both

the transmitting and receiving antenna are directional and with antenna gain pattern

shown in Fig. 3.4. Dtag−A is the distance between the tag and reader antenna A shown

in Fig. 3.3.

PositioningDtag−A

1 (m) 2 (m) 5 (m) 10 (m)

Fig. 3.9 (a) 28.7 77.3 286.2 718.3

Fig. 3.9 (b) 1.2 4.8 20.6 51.7

Fig. 3.9 (c) 1.2 4.8 20.6 51.7

Fig. 3.9 (d) 0.3 1.1 6.9 18.0

Fig. 3.9 (e) 3.6 12.2 45.4 113.8

Fig. 3.9 (f) 7.6 20.7 76.8 192.7

Fig. 3.9 (g) 0.6 2.3 12.2 30.5


The following simulations are completed following the concepts outlined in Section 3.7.

The path loss model used is shown in (3.7), with n = 2.5 for near distance and n = 3.5

for far distance. Also, the focus will be on the effect of LBT on the deployment of RFID

reader antenna which results in the Regulatory Reader Shutdown (RRS) problem. In

most of the simulation results, the -126 dBW (the LBT threshold for 2 W ERP operation

shown in Table 3.2) boundary is emboldened.

Page 52


−91.1

−86.1

−96.1 −101 −106 −111 −116 −121

−126

Figure 3.10. Simulation results on one antenna simulation. Results from simulation (vertical

and horizontal axes in m, representing the y and x axes shown in Fig. 3.7, received power

in dBW). For a second antenna with the same y coordinate to utilise the same channel

for transmission while adhering to the LBT provision, the next possible isotropic an-

tenna must be 350 m away. If using the same y coordinate but using an antenna with

side gain of approximate -14 dB, this antenna must be approximately 150 m away.

3.9.1 One Antenna Simulation

The one antenna simulation described here is similar to the example given in Sec-

tion 3.7.1. The only difference is the simulation here covers a larger area to identify

available zones to deploy a second antenna working in a same channel. From Table 3.2,

the maximum radiated power allowed by current European regulations is 2 W ERP, or

equivalently 3.2 W EIRP. From Table 3.2, the minimum threshold is -126 dBW. The

simulation result is shown in Fig. 3.10, and shows that for the next antenna to be given

the same y coordinate and to be able to operate in the same channel (assuming the

second antenna is an isotropic radiator) with the initial antenna under consideration,

the antennas must be separated by a distance of around 350 m away in the horizontal

direction.

If the second antenna (still an isotropic radiator) is to be placed in the rear position

of the first antenna, the minimum distance between them would be shorter, and is

Page 53


Table 3.4. Safe distance for different antenna positionings to avoid Regulatory Reader Shut-

down (RRS) problem. The different antenna positionings are defined in Fig. 3.9. All

these results are obtained through simulations using the simulation script described in

Section 3.7. Note that these results are for cases where both the transmitting and

receiving antenna are directional and with antenna gain pattern shown in Fig. 3.4.

Positioning Safe Distance (m)

Fig. 3.9 (a) 1900

Fig. 3.9 (b) 150

Fig. 3.9 (c) 150

Fig. 3.9 (d) 50

Fig. 3.9 (e) 300

Fig. 3.9 (f) 500

Fig. 3.9 (g) 80

approximately 210 m. This result is not obtainable from Fig. 3.10, but is from another

run of simulation, where the transmitting antenna is positioned is a suitable position

to give more view on the area behind the transmitting antenna.

However in practice, most likely the second antenna is not an isotropic radiator, and

will be the same model as the first antenna with a gain of 6 dBi. Hence the orientation

of the second antenna will be crucial to determine the minimum distance between the

first and second antennas.

A similar analysis to the analysis performed in Section 3.8.2 was carried out. Again,

referring to the antenna gain pattern shown in Fig. 3.4, the gain of the antenna in front,

behind and at the sides are, 6 dBi, -22 dBi, and -14 dBi respectively.

This gain must be considered when deciding where the second antenna can be posi-

tioned. Common positioning configurations have been shown in Fig. 3.9. The safe

distances for these configurations are obtained through simulations and are presented

in Table 3.4.

Page 54


Also, similarly to the analysis presented in Section 3.8.2, the safe distance between two

antennas with a angle between them (such as the one shown in Fig. 3.9 (h), depends

heavily on the radiation pattern, the direction of maximum radiation, and the angle,

θ. Hence, the safe distance cannot be presented here as a guideline but can be only be

obtained on a case by case basis through simulation.

3.9.2 Two Antenna Simulation

From the one antenna simulation, it has been shown that another antenna operating

in the same channel can be deployed around 350 m horizontally away from the first

antenna if the second antenna is a isotropic radiator. In practice, where a directional

6 dBi antenna is used (Fig. 3.4), the second antenna can be placed 150 m horizontally

away from the first antenna, as is indicated in Table 3.4.

In the two antenna simulation discussed here, a second antenna is deployed 150 m

horizontally away from the first one, and the result is as shown in Fig. 3.11. The bold

line is the approximate safe boundary for the deployment of an isotropic radiator in the

same channel. The distance is approaching 1.6 km in front of both antennas and around

400 m horizontally away from the second antenna. If a directional 6 dBi antenna is

used as the third antenna (which is most likely the case in reality), the safe distance

depends on the positioning and orientation of the third antenna. If the placement of

third antenna takes the form of Fig. 3.9 (a), facing directly to either of the first two

antenna, it has be to placed approximately 2.2 km away (outside the solution space

shown in Fig. 3.11). If the third antenna is to be positioned side-by-side with the first

two antennas, the safe distance will be around 150 m away from the nearest antenna

(well within the solution space shown in Fig. 3.11).

This shows that for antennas transmitting at the same frequency even with care in

orientation, they have to be at a great distance apart to avoid RRS problem and hence

this deployment is not practical in real life applications where LBT is enforced.

Page 55


−129

−126

−124

−121

Figure 3.11. Simulation results of a two antenna simulation. Results from simulation (vertical

and horizontal axes in units of 10 m, representing the y and x axes shown in Fig. 3.7,

received power in dBW). The two antennas are positioned at locations (50,50) and

(200,50) respectively. The bold line is the approximate safe boundary for the deploy-

ment of an isotropic radiator in the same channel.

3.9.3 One Antenna in Horizontal Position

Another antenna placement configuration is to place an antenna in a horizontal po-

sition, with the direction of maximum gain perpendicular to the ground plane. An

antenna in this fashion can have the interrogation field projected upwards or down-

wards, as shown in Fig. 3.12. Whenever the term horizontal antenna is mentioned

throughout this chapter, it is referred to the configuration as illustrated in Fig. 3.12

with a certain elevation, which will need to be specified either as a default value of 0 m

or as a separately specified value that overrides the default.

An antenna in a horizontal position can be on any elevation level with respect to the

ground. A UHF reader antenna is usually a patch antenna and the patch will be parallel

with the ground and may be radiating towards the ground. However, the developed

simulation software does not consider the reflection from the ground. Hence, the de-

veloped software is more accurate in the case where there is no antenna facing towards

ground.

Page 56


1 m

1 m

Default AntennaConfiguration

Ground

Horizontal Antenna(0 m Elevation, Upwards)

Horizontal Antenna(2 m Elevation, Downwards)

Figure 3.12. The configuration of a horizontal antenna on ground or facing the ground with

respect to the normal configuration of a transmitting antenna. The simulation

software does not consider the reflection from the ground in the case where an antenna

is facing towards the ground. The thickness of a reader antenna is not considered.

The MATLAB software developed described in Section 3.7 was updated to handle hori-

zontal antennas with any elevation. One additional input parameter is required, which

is the elevation level with respect to a reference plane. The default reference plane is

the x-y plane 1 m above the ground shown in Fig. 3.7.

By replacing the rightmost antenna in Fig. 3.11 with a horizontal antenna (0 m eleva-

tion), the simulation result, shown in Fig. 3.13 when compared with Fig. 3.11, shows

that the boundary to deploy a third reader antenna operating in the same frequency

channel moves slightly closer to the first and second antennas, which improves the

situation slightly.

3.9.4 Antennas Operating in Different Channels

All the discussions on the Regulatory Reader Shutdown (RRS) problem presented in

previous sections are on multiple reader antennas operating in the same channel. In

this section, the effect of readers operating in neighbouring channels will be investi-

gated in the context of RRS.

Page 57


−124

−126

−129

−131

−133 −136 −138 −140 −143

−121

−119

−117

Figure 3.13. Simulation results showing the effect of horizontal antenna configuration. Re-

sults from simulation (vertical and horizontal axes in units of 10 m, representing the y

and x axes shown in Fig. 3.7, received power in dBW). The bold line is the approximate

safe boundary of the deployment of an isotropic radiator in a same channel following

LBT. This figure shows that mixture of antennas with different orientation will improve

the dense reader environment as compared to Fig. 3.11.

For the case of a multiple-interrogator environment (Section 3.6.3), a simple and con-

servative way to give a rough estimation of inter-channel interference is available if all

the readers are compliant to [35]. Referring to the transmit mask shown in Fig. 3.5(a), it

can be seen that if a channel is being used for interrogation, the leakage to its immedi-

ate neighbouring channels must be below 20 dBch. Hence, if the simulation is focused

on a particular channel, the interference from the next neighbouring channels can be

emulated by reducing the signal strength by 20 dB.

However, it is trickier for the case of a dense-interrogator environment. Referring to

Fig. 3.5(b), the transmit mask depends on Tari, which is the reference time interval for

a data-0 in Interrogator-to-Tag signalling, and is between 6.25 µs and 25 µs [35]. In the

case study presented in this section, the focus will be on the European dense reader

environment, where the bandwidth of a channel is 200 kHz. Hence, for simplicity, Tari

is chosen to be 12.5 µs. By doing this, the -30 dBch level in dense-interrogator mask

Page 58


shown in Fig. 3.5(b) will coincide with the immediate neighbouring channels and the

-60 dBch level will coincide with the channels two channels away from the channel of

interest.

It must be noted that these transmit masks provided by [35] are international guide-

lines and are superseded by local regulations if the local regulations impose a stricter

transmit mask. This may be the case in a European country where ETSI EN 302 208 [48]

is used as the local regulation. However, at the time when this research was carried

out, [48] was undergoing a number of changes. Hence it was decided to use the trans-

mit mask for a dense-interrogator environment from [35] to demonstrate the ability of

the developed software to handle simulations involving antennas operating in differ-

ent channels.

An example to demonstrate the use of the developed simulation software to handle

the case where antennas are operating in different channels is shown in Fig. 3.14. The

antenna placement in this example is the same as the case shown in Fig. 3.13, where

there are a total of two transmitting antennas, located at (50,50) and (200,50). The

antenna located at (50,50) is of normal orientation with maximum gain facing in the y

axis, while the antenna located at (200,50) is of horizontal orientation with maximum

gain facing the z axis and has a 0 m elevation (refer to Fig. 3.12). Note that the difference

in height between the horizontal antenna and the normal configuration antenna has

been taken into account in the calculation. Also, in this scenario, these two antennas

are operating in the immediate neighbouring channels of the channel interest. Hence

to identify the safe zone to deploy a third transmitting antenna (which is an isotropic

antenna for simplicity), which is to operate in the channel of interest (not the channel(s)

occupied by antennas at (50,50) and (200,50)), the signals from antennas at (50,50) and

(200,50) are reduced by 30 dB, following the transmit mask for a dense-interrogator

environment shown in Fig. 3.5.

To avoid confusion in forming a guideline, the two antennas shown in Fig. 3.14 are re-

simulated individually and included as Fig. 3.15 to identify the safe distance between

a first transmitting antenna with gain pattern shown in Fig. 3.4 and a second antenna

(an isotropic radiator) to adhere the LBT provision. Fig. 3.15(a) shows the simulation

Page 59


−126

−128

−130

−132

−134 −136 −138 −140 −142 −144 −146

−124

−122

−126

−124

Figure 3.14. Simulation results for some antennas operating in different channels and point-

ing in different directions. Results from simulation (vertical and horizontal axes in

m, representing the y and x axes shown in Fig. 3.7, received power in dBW). Transmit

mask for dense-interrogator environment is used with Tari chosen to be 12.5 µs. Note

that both antennas are operating in an immediate neighbouring channel of the channel

of interest.

results of an antenna with normal configuration while Fig. 3.15(a) shows the simulation

results of an antenna with horizontal configuration with 0 m elevation. Note that for

the case of an antenna with horizontal configuration, the simulation x-y plane is with

an elevation of 1 m (refer to Fig. 3.7 for an illustration of the position of the x-y plane).

Table 3.5 shows the safe distances between a first transmitting antenna with gain pat-

tern shown in Fig. 3.4 and a second antenna (an isotropic radiator) for various sep-

aration between the channel of interest which will be used by the newly introduced

antenna and the channels used by the first (existing) antenna. For a transmission oper-

ation in the channel of interest, the antenna must be placed 50 m away in all directions

from a horizontal antenna operating in the immediately neighbouring channel. For

an antenna in the default configuration (Refer Fig. 3.12) operating in the immediately

neighbouring channel, the distance is around 180 m from the front, 45 m from both

sides and 30 m from the back.

Page 60


−123

−126

−130

−134 −138 −142

−119

(a) First antenna in normal configuration

−126

−133

−138

−143

−148

−153

(b) First antenna in horizontal configuration

Figure 3.15. Determination of safe distance in accordance to LBT when the second antenna

is an isotropic radiator. Results from simulation (vertical and horizontal axes in m,

representing the y and x axes shown in Fig. 3.7, received power in dBW). Results for

both antenna in normal and horizontal configurations are tabulated in Table 3.5. Note

that both antennas in (a) and (b) are operating in an immediate neighbouring channel

of the channel of interest.

Page 61


If the second antenna is not an isotropic radiator but is instead an antenna with similar

gain pattern to the first antenna, the safe distance will then depend on the orientation

and placement of both antennas. Nonetheless, the safe distance can be easily deduced

from the graphical presentation of the results such as those presented in Fig. 3.15.

For example, to locate a 6 dBi antenna facing the horizontal antenna at position (200,50)

in Fig. 3.15(b) and with channel difference of 1, the boundary must be as low as -

132 dBW (-126dBW-6dBi). From Fig. 3.15(b), the distance from the -132 dBW boundary

to antenna at (200,50) is approximately 75 m. Note that the safe distance is 50 m when

the second antenna is just an isotropic radiator and not a 6 dBi antenna.

Table 3.5 is sufficient to serve as a guideline on antenna placement in adherence to

the LBT provision in an RFID deployment zone. The only restriction is that all the

readers must be able to be pre-programmed to operate only in a pre-selected frequency

channel. However, upon a more careful inspection, it is not hard to realise that the safe

distances shown in Table 3.5 are unreasonably large. Hence the enforcement of LBT

will bring unreasonable restrictions to RFID deployment.

Table 3.5. Safe distance for different antenna configuration when the second antenna is an

isotropic radiator in accordance to LBT. Distances for channel difference 1 obtained

from Fig. 3.15(a) and Fig. 3.15(b). Results for other cases of channel difference were

obtained through separate simulations.

Channel DifferenceAntenna (Normal Configuration) Horizontal Antenna

Front (m) Side (m) Back (m) (m)

0 1400 350 210 320

1 180 45 30 50

2 130 25 15 35

3 95 20 10 30

Page 62


3.9.5 A More Hostile Environment

All the previous simulations are performed based on the path loss model as shown in

(3.7), using n = 2.5 for near distance and n = 3.5 for far distance. Hence, the simula-

tion results are approximate indications of various real life situations. A more hostile

environment will have a higher n and the safe distances in Table 3.5 will be reduced.

On the other hand, a less hostile environment will have increased distances. However,

it does not mean that RFID will perform better in a hostile environment. It is just that

in a dense reader and hostile environment, reader placement would be less restricted

under the “Listen Before Talk” provision as compared with that for a less hostile envi-

ronment.

3.10 Real Life Application

This section explores some real life applications where the coverage of an RFID in-

terrogation zone can be improved. Similarly to Section 3.9, the focus will be on the

Regulatory Reader Shutdown (RRS) problem, though the same idea can applied to

reduce the Reader Interference with Tag Replies (RITR) problem at the same time.

3.10.1 A Checkout Counter

As mentioned in Section 3.9.3, the term “horizontal antenna” is referred to the config-

uration as illustrated in Fig. 3.12 with certain elevation (default of 0 m). However, it

is usually not feasible to place a horizontal antenna directly on the ground, except in

some very specific applications. For example, in checkout points of retail outlets, it is

more sensible to place the antenna on the checkout counter level rather than ground

level (Fig. 3.16). The simulation results presented for the horizontal antenna are still

valid at far distance in all directions. There will be a slight difference in the results at

near distance. Alternatively, since the simulation software can handle a horizontal an-

tenna with any elevation, a new set of simulation results could provide a more accurate

prediction of the signal strength as received by an isotropic antenna at any location of

interest relative to checkout counter.

Page 63

3.10 Real Life Application

1 m

Ground


CheckoutItems


Figure 3.16. Repositioning of a horizontal antenna. Lifting a horizontal reader on the ground

to counter level.

3.10.2 Reader Synchronisation

One simple method to solve the RRS problem and adhere to the European regulation

(LBT provision) at the same time is to synchronise all the readers. All the readers will

be caused to start to “Listen” at the same time, and assuming no other short range

device is operating, will detect no signal at all as no reader is transmitting signals.

Following the European regulation, if no signal is detected in the “Listen” period, the

“Listen” period will end at the same time for all the readers [48], and the readers can

start interrogating tags, as shown in Fig. 3.17. Using reader synchronisation, readers

can avoid the RRS problem caused by LBT completely. A complete treatment of reader

synchronisation can be found in Chapter 5.

However it should be noted that reader synchronisation does not prevent other forms

of collision problem in a dense reader environment, such as the Reader Interference

with Tag Replies (RITR) problem. To minimise the RITR problem (Section 3.5), proper

separation and orientation of reader antennas are required. For simple cases, Table 3.3

can be used as a guideline for placing readers. For more complex cases, simulations

are required for each individual cases.

Page 64


Reader 1

Reader 2

"Listen" "Talk"

"Listen"

Reader N "Listen"

"Talk"

"Talk"

Time

Figure 3.17. Reader synchronisation. All the reader antennas are preprogrammed to “Listen” at

the same time before entering the “Talk” stage.

3.11 Conclusion

A detailed analysis of RFID indoor propagation models has been performed and dis-

cussed. This effort aimed to create a comprehensive guide to allow fast and successful

deployment of large-scale RFID systems and to maximise the potential and benefits of

that deployment. This path loss model can be applied to predict the signal strength

of an RFID interrogation signal at a certain distance away from a transmitting an-

tenna. The model has been implemented using a simulation script written in MAT-

LAB. Exemplary simulation results were presented, together with some insights and

discussions. Safe distances between antennas were suggested and these results can be

used in RFID large-scale deployment areas to avoid the problem of Reader Interfer-

ence with Tag Replies (RITR) and the problem of Regulatory Reader Shutdown (RRS)

in European countries with “Listen Before Talk” (LBT) provision. It has been shown

that strict regulations (such as LBT) will in fact limit large-scale deployment of RFID

systems. Proper antenna positioning has been proven to be effective in minimising

the problem of RITR. Reader synchronisation has been introduced to avoid the RRS

problem (a complete treatment of reader synchronisation can be found in Chapter 5).

More field testing results can be found in Chapter 4, where discussions on operational

considerations in simulation and deployment of RFID systems are presented.

Special Note: The research presented in this chapter was carried out in year 2005, with findings

showing the provision of “Listen Before Talk” (LBT) will impede large-scale deployment of

Page 65

3.11 Conclusion

RFID systems. LBT is officially removed from the latest version of ETSI 302 208 in year

2007 [59].

Page 66

Chapter 4

RFID OperationalConsiderations

LARGE-scale radio frequency identification (RFID) deployment

is needed for efficient item identification in supply chains. To re-

duce cost and save time, simulations are often carried out before

actual implementation, especially when RFID is used in regions in which

strict regulations and standards must be adhered to. However, due to the

unpredictable environmental effects on radio propagation, simple simula-

tion results can be misleading and questions have been raised over the va-

lidity of many wireless simulations. This research study reviews, from the

point of RFID antenna deployment, the sources of error in wireless simula-

tions reported in some publications. Also, the idea of second carrier sensing

is investigated to reduce the Reader Induced Tag Confusion (RITC) prob-

lem, and Specific Absorption Rate (SAR) of a dense reader deployment is

discussed to ensure human safety.

Page 67

4.1 Introduction

4.1 Introduction

This chapter is based extensively on the foundation work presented in Chapter 3.

There will be some overlapping of material between this chapter and Chapter 3. This

is to avoid extensive cross-chapter referencing, but an effort has been made to keep the

repetition to a minimal level.

In Chapter 3, a path loss model was proposed to simulate a dense reader environment.

A simulation software was developed to provide guidelines in antenna placements.

The main aims were to reduce two collision problems in a dense reader environment,

namely the Reader Interference with Tag Replies (RITR) problem and the Regulatory

Reader Shutdown (RRS) problem.

A good simulation of RFID deployment is essential as such simulation enables visu-

alisation of RFID deployment prior to actual implementation. However, a simulation

result depends on the complexity of the model used for the simulation. A simple simu-

lation can often lead to incorrect conclusions while a complex simulation may consume

too much computational time to be feasible. Several papers in the literature have in-

vestigated the validity and credibility of wireless network simulations [60, 61, 62]. Of

particular interest is [60], where several common simulation assumptions, which con-

tribute the most to simulation inaccuracy, are identified.

In this chapter, common simulation errors, such as use of an inappropriate simula-

tion model, neglect of antenna gain, and misinterpretation of simulation results, will

be discussed in the context of RFID, along with the exploration of the challenges in

RFID simulations. The simulation software used in this chapter is the same simulation

software presented in Section 3.7. Simulation results obtained through this developed

software are compared with measurement results, and several suggestions have been

offered to minimise the possible sources of error.

This chapter is structured in the following way. At first, RFID EMC related background

information is discussed in the next section, followed in Section 4.3 by the brief intro-

duction of the EPCglobal C1G2 RFID protocol and its proposed transmit mask in a

dense reader environment. As mentioned before, Section 4.2 and Section 4.3 contain

Page 68

Chapter 4 RFID Operational Considerations

some overlapping material with Chapter 3, but this is to enhance readability of this

chapter and reduce the frequency of cross-chapter referencing.

Section 4.4 presents a path loss experiment while Section 4.5 looks into several com-

mon sources of error in wireless simulation, and provides suggestions and ways of

minimising them.

Besides the discussion of sources of simulation error, this chapter also investigates two

further matters relevant to RFID operational considerations. Section 4.6 offers some

insight into the idea of second carrier sensing in reducing the Reader Induced Tag

Confusion (RITC) problem, its threshold limit selection and its dependency on an-

tenna positioning while Section 4.7 investigates the Specific Absorption Rate (SAR),

required to ensure human safety in a dense reader environment. Section 4.8 provides

the conclusions of this chapter.

Early parts of this chapter up to Section 4.5 have been published in [63]. Section 4.6 has

appeared in conference as [64].

4.2 RFID EMC Background

The regulatory status for using RFID in the UHF spectrum around the world can be

summarised into two categories: governed under Frequency Hopping Spread Spec-

trum (FHSS), and governed under Listen Before Talk (LBT). By the end of 2005, it is

expected that 50 countries, representing 83 % of the global GNI (Gross National In-

come) will have RFID regulations in the UHF band [65]. It is essential to understand

the differences between each regulation before an RFID simulation on deployment is

carried out, especially on the allocated bandwidth and the maximum allowable radi-

ated power.

Page 69

4.3 RFID Protocol

4.2.1 Frequency Hopping Spread Spectrum (FHSS)

An example under this category is the USA FCC Title 47 Part 15.247, with operation

within the band 902-928 MHz [47]. This bandwidth is then subdivided into 50 chan-

nels of 500 kHz. Each channel has a minimum transmission time of 0.4 s. Maximum

transmit power is 1 W with a maximum of antenna gain of 6 dBi, giving a maximum

Effective Isotropic Radiated Power (EIRP) of 4 W. It is adopted mainly in North and

South America.

4.2.2 Listen Before Talk (LBT)

An example under this category is the European ETSI 302 208 [48] that has been adopted

by some European countries. It allocates the frequency band of 865 - 868 MHz for RFID

deployment. This frequency band is then divided into 15 sub-bands or channels; each

spans a total of 200 kHz. When a reader is operating at the maximum total radiated

power, which is 2 W ERP (Effective Radiated Power) or equivalently to 3.2 W EIRP,

only 10 sub-bands are available, while the remaining 5 are utilised as guard bands in

which a low ERP is allowed. Three guard bands (out of the total five) are located at the

lower end of the designated bandwidth with a maximum ERP of 100 mW and the two

remaining guards bands are located at the higher end of the designated bandwidth

with a maximum ERP of 500 mW.

The detailed discussion of LBT can be found in Section 3.6.4. The transmit power and

the corresponding threshold values are extracted from the above-mentioned ETSI doc-

ument and integrated into Table 3.2. It is reproduced as Table 4.1 here for convenient

reference.

4.3 RFID Protocol

The operation of an RFID system is also standardised to encourage wide spread de-

ployment. EPCglobal has produced ”EPC Radio-frequency Identity Protocols Class-1

Generation-2 UHF RFID Protocol for Communication at 860 - 960 MHz” [35], in short

Page 70


Table 4.1. UHF RFID reader effective radiated power with corresponding threshold values

for LBT. Threshold value obtained from [48]. (From Table 3.2.)

ERP (W) ERP (dBW) Threshold (dBW)

Up to 0.1 Up to -10 ≤-113

0.1 to 0.5 -10 to -3 ≤-120

0.5 to 2.0 -3 to 3 ≤-126

EPC C1G2, as the standard operational protocol. This comprehensive protocol includes

the detailed air interface between readers and tags and also standard commands for tag

interrogation. The section in the protocol which is closely related to this research is on

the transmit mask for dense-interrogator environments as shown as Fig. 3.5. A more

relaxed transmit mask is used for low reader density surroundings [35] but is not pre-

sented here, as the focus of this research will be in a dense reader environment. For a

narrow bandwidth channel (European 200 kHz channel), it is suggested in this protocol

that odd-numbered channels should be used for tag backscatter while even-numbered

channels will be used for reader interrogation signals. For a wide bandwidth channel

(USA FCC 500 KHz channel), all available channels can be used for reader interroga-

tion, as tag backscatter signals will be located at the boundaries of these channels.

4.4 Path Loss Measurement Experiment

A simple path loss measurement experiment was described in outline in Section 3.3.

This section presents more details on the equipment setup, experiment locations and

data analysis.

The main objectives of this experiment are:

1. To demonstrate the multi-path effect on in-building path loss.

2. To provide experimental data for the investigation of in-building path loss model,

and to suggest the best model for a particular room.

Page 71


4.4.1 Experimental Setup:

The equipment used during the experiment is shown in Fig. 4.1. The antennas used

are approximately circularly polarized and have a gain of 8 dBi. In this experiment, we

are using a signal with a low transmit power generated by the HP ESG-3000A signal

generator, and hence the transmitted signal is in compliance with regulations.

Much of the theory concerning reflections which can be found in Sections 4.4.3 and

4.4.4 assumes linear polarisation. For direct rays between the two antennas the circular

polarisation produced by the test antennas will suffer the same loss as would linear

polarisation. However circular polarisation produced by these antennas would have

its sense reversed on reflection from metallic surfaces, and should not then be detected

by the receiving antenna.

Although at first sight this may seem to be a significant error for those paths involving

reflection, the polarisation produced by the antennas is imperfect even on bore sight,

and will be highly elliptical at the radiation angles involved in many reflections. It will

be particularly so for those paths for which the ray tracing of Section 4.4.4 is performed.

Moreover, most of the measurements in our path loss measurements involve direct

rays.

So we will regard the analyses and measurements in Sections 4.4.3 to 4.5.1 as capable

of providing useful data on which path loss models may be constructed.

4.4.2 Room Grid

Measurements were carried on a specified grid as shown in Fig. 4.2. The grid was

created in Room N203, Engineering North, the University of Adelaide, North Ter-

race Campus, South Australia. This room contains metallic cupboards, drawers, and

wooden tables and is chosen as an experimental site to represent a typical storage area.

The room is divided into a grid system, with markers positioned 1 m apart. Two an-

tennas of known gain are used as the transmitter and receiver. These two antennas

are directly facing each other (with each at exactly the same height) in order to obtain

Page 72


(a) Cushcraft 8 dBi S9028PC directional antenna.

(b) HP 8594E spectrum analyser

9 kHz - 2.9 GHz

(c) HP ESG-3000A signal generator

250 kHz - 3000 MHz

Figure 4.1. Equipment used in path loss prediction experiment.

for each an orientation with maximum directivity. Several measurements are taken at

various positions for a chosen separation and averaged.

4.4.3 Experimental Results

The raw data obtained from the experiment is shown in Table 4.2. The explanation

corresponding to each column is as follow:

• r: The distance between the transmitting and receiving antennas.

• Tx-Pt: The location of the transmitting antenna with reference to the room grid

as shown in Fig. 4.2.

Page 73


Figure 4.2. Experiment grid for signal strength measurement. For measurement more than 8 m,

the receiver antenna is moved out of the room through the door shown. The antennas

are facing each other with maximum gain attained for all measurements. Each grid is

1 m × 1 m.

• Rx-Pt: The location of the receiving antenna with reference to the room grid as

shown in Fig. 4.2.

• Tx-Pw: The transmitted power (in dBm) as delivered by HP ESG-3000 signal

generator (Fig. 4.1(c)). The power level is set, prior to the experiment, to -20 dBm

and is used for the entire measuring session.

• Rx-Pw: The received power (in dBm) as measured by HP 8594E spectrum anal-

yser (Fig. 4.1(b)).

• Gain: The total antenna gain (in dBi) of both the transmitting and receiving an-

tenna (Fig. 4.1(a)) used in the experiment.

• Loss: The total losses caused by coaxial cables used in connecting antenna (Fig. 4.1(a))

to spectrum analyser (Fig. 4.1(b)) and antenna (Fig. 4.1(a)) to signal generator

(Fig. 4.1(c)) . The loss value presented in Table 4.2 is an assumed value (in dB).

• PL: The path loss (PL), in dB, calculated using the measured power, taking con-

sideration of antenna gain and cable loss: PL = (Rx-Pw) - (Tx-Pw) + (Gain) -

(Loss), where Rx-Pw, Tx-Pw, Gain and Loss are defined above.

• FSPL: Theoretical free space path loss (FSPL), in dB, of the distance r, calculated

using

FSPL =(

λ

4πr

)2

(4.1)

Page 74


• Diff: The difference between the PL and FSPL, calculated using (PL) - (FSPL).

• Ave: The average of the Diff for all the same r. For example, a total of six mea-

surements were taken at different locations for the case of r = 1. The six Diff

measurement results were averaged to obtain an Ave value. This Ave value is the

same for all the cases with the same r. For the cases where r = 7 and 8, as there is

less than four measurement results, “N/A” (indicating not available) was input

in the Ave column instead.

Table 4.2. Average prediction of path loss based on experimental results. Explanation of

column titles can be found in Section 4.4.3, with Tx-Pw = -20dBm; Gain = 16dBi; and

Loss = 3dB.

r Tx-Pt Rx-Pt Rx-Pw PL FSPL Diff Ave

1 B8 C8 -39.5 32.5 31.67 0.83 0.95

1 A8 B8 -39.3 32.3 31.67 0.63 0.95

1 B9 B8 -40.0 33.0 31.67 1.33 0.95

1 C9 C8 -39.3 32.3 31.67 0.63 0.95

1 A6 A7 -39.6 32.6 31.67 0.93 0.95

1 D7 D6 -40.0 33.0 31.67 1.33 0.95

2 A8 C8 -45.5 38.5 37.69 0.81 0.39

2 B9 B7 -46.7 39.7 37.69 2.01 0.39

2 C9 C7 -46.0 39.0 37.69 1.31 0.39

2 A6 A8 -42.7 35.7 37.69 -1.99 0.39

2 D7 D5 -44.5 37.5 37.69 -0.19 0.39

3 A9 A6 -46.8 39.8 41.21 -1.41 -1.07

3 D7 A7 -48.8 41.8 41.21 0.59 -1.07

3 A6 A9 -45.5 38.5 41.21 -2.71 -1.07

3 D7 D4 -46.8 39.8 41.21 -1.41 -1.07

3 C7 C4 -47.8 40.8 41.21 -0.41 -1.07

4 D9 D5 -50.0 43.0 43.71 -0.71 0.44

Continued on next page

Page 75


Table 4.2 – Continued from previous page

r Tx-Pt Rx-Pt Rx-Pw PL FSPL Diff Ave

4 C9 C5 -51.6 44.6 43.71 0.89 0.44

4 C7 C3 -53.9 46.9 43.71 3.19 0.44

4 A9 A5 -49.1 42.1 43.71 -1.61 0.44

5 D9 D4 -48.8 42.8 45.65 -2.85 -1.28

5 B9 B4 -52.5 45.5 45.65 -0.15 -1.28

5 A9 A4 -49.3 42.3 45.65 -3.35 -1.28

5 C7 C2 -53.9 46.9 45.65 1.25 -1.28

6 D9 D3 -53.0 46.0 47.23 -1.23 -0.03

6 A9 A3 -53.3 46.3 47.23 -0.93 -0.03

6 C8 C2 -55.5 48.5 47.23 1.27 -0.03

6 B9 B3 -55.0 48.0 47.23 0.77 -0.03

7 A9 A2 -55.4 48.4 48.57 -0.17 N/A

7 C9 C2 -56.8 49.8 48.57 1.23 N/A

7 B9 B2 -59.0 52.0 48.57 3.43 N/A

8 A9 A1 -61.0 54.0 49.73 4.27 N/A

4.4.4 Discussion

1. It is suspected that PL at all near distances would be slightly lower than the FSPL.

This may be due to the multi-path effect where the propagating waves transmit-

ted by the antenna are reflected by the surrounding objects. At certain locations,

constructive interference will take place and will result in a detection of a stronger

received signal as compared with a received signal in a reflection free environ-

ment (such as an anechoic chamber). However, it has to be noted that destructive

interference may also occur.

To illustrate this effect, we can consider the addition of two waves, with different

phase. The power expression is as shown below:

Page 76


Preceived = (acosβ1z + bcosβ2z)2

= a2cos2β1z + b2cos2β2z + 2abcosβ1zcosβ2z

If a = 1, b can have among others the values of 1, j,−1,−j. The |Preceived| is then

4, 2, 0, 2, giving an average Preceived of 2.

A different approach is to compute the expected value of Preceived as shown below:

E [Preceived] = E[(acosβ1z + bcosβ2z)2

]= E

[a2cos2β1z + b2cos2β2z + 2abcosβ1zcosβ2z

]= E

[a2cos2β1z

]+ E

[b2cos2β2z

]+E [2abcosβ1zcosβ2z]

=a2

2+

b2

2+ 2ab

12

E [cos(β1z − β2z) + cos(β1z + β2z)]

=a2

2+

b2

2

The first approach predicts that although there appears to be nulls due to reflec-

tions, on average a gain is expected. The second approach predicts on average

there will be a gain due to reflection.

2. From the Diff column in Table 4.2, it can be seen that at some distance away the

PL computed using the measured signal strength is constantly lower than the

FSPL. For example, when r is 3 m.

Figure 4.3 shows the measuring configuration from a side view. For simplicity,

only one reflected wave is considered, which is the reflection from the ground.

The reflection from the ceiling is weaker than the reflection from the ground be-

cause the measurements were carried out at an elevation of 1 m as shown in

Fig. 4.3. The room where the measurement took place is approximately 3 m in

height and the reflection path from ceiling is much longer than the reflection

path from the ground. However, reflections from the surrounding objects, such

Page 77


as metal cupboards, walls or metal shelves, can have large impact on the receiv-

ing signal, depending on the material of the objects and the path of reflection.

Although these reflections from the surrounding objects are neglected in the dis-

cussion below, inconsistency in measurements result (as seen in Table 4.2) is most

likely contributed by them.

1 m

r

D

21

Ground

2D2

Figure 4.3. Measuring configuration from side view. r is the distance between the transmitting

and receiving antenna, D is the total length of the reflection path.

Another assumption is that the angles θ1 and θ2 in Fig. 4.3 are the same. This

will not be the case if the ground surface is not flat or uneven. Again, for sim-

plicity and to have a predictable reflection path length, a flat and even ground is

assumed and hence those angles are the same.

The results are shown in Table 4.3. It is very possible that the difference between

PL and FSPL is dependent on the difference in wavelength between the incident

path length and reflected path length. A difference of half a wavelength between

the incident path length and reflected path length, after subtraction of full wave-

lengths will most probably result in higher PL, while a difference of 0 wavelength

will result in lower PL.

For example, for r = 3 or r = 5, the difference in wavelength is between 0.1 and

0.2, and hence the computed PL values based on experiment measurements are

on average lower than the FSPL. On the other hand, for r = 2 or r = 4, the differ-

ence in wavelength is between 0.4 and 0.6 and the computed PL values based on

experiment measurements are on average higher then the FSPL.

Page 78


Table 4.3. Difference in wavelength between direct distance, r, and reflected path, D. An-

tennas are mounted following Fig. 4.3. The signal generator is transmitting a signal at

915 MHz with a wavelength of approximately 0.328 m. r is the distance between the

transmitting and receiving antennas in metre, D is the length of the reflection path in me-

tre, D-r is the difference between D and r in metre, and wl is D-r in terms of wavelength,

Ave is from Table 4.2 and is the average Diff values as explained in Section 4.4.3.

r D D-r wl Ave

1 2.236 1.236 3.768 0.95

2 2.828 0.828 2.527 0.39

3 3.606 0.606 1.848 -1.07

4 4.472 0.472 1.433 0.44

5 5.385 0.385 1.173 -1.28

6 6.325 0.325 0.991 -0.03

However, this argument is not strongly supported by measurements for r = 6.

From Table 4.3 and based on the pattern shown by r = 2, 3, 4 and 5, as explained

above, the PL values should be lower than FSPL. This inconsistency is most prob-

ably caused by reflection from nearby objects. For r greater than 6 m, too few

measurements are taken to form any conclusion. For a measurement more than

8 m, the transmitting and receiving antennas are separated by walls and the ef-

fect of walls must be taken into account. It is assumed that walls will decrease

the measured signal strength and hence increase the computed PL value.

This experiment has achieved objective 1 but has yet to produce a model, which

can best describe the observation of multi-path effects, mainly caused by reflec-

tion. More measurements were carried out in different locations (Appendix F) to

provide further understanding and insight in this study.

Page 79

4.5 Sources of Simulation Error


A good simulation program should be able take into account the EMC regulations and

protocol standards discussed in Section 4.2 and Section 4.3. Also, the common sources

of simulation error of a wireless system as outlined in [60] and [61] must be minimised.

4.5.1 Path Loss Model

The term “path loss” is widely used in the literature and is carefully explained by the

authors in Section 3.2 and in Ref. [56]. Often the path loss model in a simulation is con-

sidered as a simple function of distance. This is only true for free space propagation,

and is used for satellite communication simulations [58]. A typical RFID deployment

zone is a warehouse filled with commercial products and logically a more complex

model is required. Hashemi [66] has categorised all the path loss models found in the

literature, apart from the simple free space path loss, into four groups.

These four categories have been studied carefully. It is found that a path loss model

with variable environmental factor, n, is suitable for the case of RFID as discussed in

Section 3.2, and this path loss model is reproduced here for convenient referencing:

PL(dB) = PL(d0) + 10 × n × log10

(dd0

)(4.2)

where d0 is an arbitrary reference distance, n is the environment factor, d is the separa-

tion distance between two antennas and PL(d0) is the free space path loss for a distance

d0. The n value takes into consideration, path obstruction, reflection, absorption and

other attenuation effects introduced by the presence of objects inside a building.

Also discussed previously in Section 3.2, (4.2) does not consider the fact that the n

value will increase as the distance increases. The modified version of (4.2) is presented

in Section 3.4 and is reproduced here as (4.3):

Page 80


0

20

40

60

80

100

120

1 10 100Distance (m)

Pa

th L

oss

(d

B)

n=2

n=3

n=4

Eq. (5.3)

Measured

1st Wall Encontered

More Walls

Encontered

Figure 4.4. Plot of path loss using different models together with measured results..

PL(dB) =

⎧⎪⎪⎪⎨⎪⎪⎪⎩

PL(d0) + 10 × n1 × log10dd0

0 < d ≤ 8 m

C + 10 × n2 × log10dd0

d ≥ 8 m

(4.3)

where at 915 MHz with n1 = 2.5 and n2 = 3.5, PL(d0) = 32 dB and C = 22.97 dB (see

Section 3.4 for details).

The comparison with various models and experimental results is plotted in Fig. 4.4

that shows a path loss model in free space, three models based on (4.2) with n = 2, 3 or

4, one model based on (4.3) and also the practical measurements from experiment. It

can be seen that (4.2), with n = 2 and PL(d0) set to the value for free space, is actually

the same as free space path loss. As the n value increases, the path loss increases for a

fixed distance.

The practical measurement was set up as shown in Fig. 4.2 in a room as described

in Section 4.4.2. The measurement shows several things. The first is that within the

room, the path loss is less than the free space path loss generally by a few dB. This

we attribute to reflections within the room, which would be expected to produce this

effect. Secondly, the practical measurement results follow an approximately free space

path loss pattern until the first wall is encountered along the measuring path. As the

Page 81


distance increases, more major obstacles appear along the path of measurement, and

the practical measurement results have shown path losses higher than a free space path

loss. Equation (4.3) best mimics the measured results. Although the error margin can

be up to 10 dB, it can be minimised, if more appropriate values of n1 and n2 are used

in (4.3). However, the n value changes from location to location. On-site measurement

must be carried out to determine the best n for that certain area before the values of n

are used in simulations. Walls have been shown to have a great impact on the value of

n. Within a room, one fixed n value can be used, but n must be increased when a wall is

encountered as the distance increases. Equation (4.3) offers quick path loss estimation.

More complex surroundings with more obstacles in nearby vicinity requires a better

model as discussed in next section.

4.5.2 Reflection, Refraction, and Diffraction

A path loss model with environmentally determined n (such as (4.3)) has a high level

of inaccuracy if the environment of interest is full of obstacles, such as walls, metallic

cupboards, and narrow corridors. For example, in Fig. 4.5, the path loss in (a) and

(b) will not be the same, and will not be predicted by (4.3). This is caused by wave

propagation effects such as reflection, refraction and diffraction. Also, the appearance

of obstacles, such as a wall, must be accounted for the calculation of path loss.

The ray tracing method is well explained in the literature, such as [67, 68]. It is based on

geometrical optics and complemented by geometrical theory of diffraction to take into

consideration reflection, refraction and diffraction. On the other hand, [69, 70] modify

(4.2) empirically to cater for reflection, refraction and diffraction.

We combined both of the above ideas to produce a hybrid path loss model. The ar-

gument is that ray tracing normally requires many traces. If we limit ourselves to not

more than 5 rays, with each ray using an empirical model [69, 70], it is hoped that a

simple and accurate model can be obtained.

Using the settings presented in Fig. 4.5 (b), we constructed two additional rays (P1 and

P2) apart from the direct ray from transmitting antenna (Tx) to the receiver antenna

Page 82


(a)

Ceiling

(b)

Tx Rx Tx Rx

Ground Ground

P

P

1

2

MetallicBox

Figure 4.5. Comparison of the received signal strength with and without the presence of a

nearby metallic box. Signal received by receiver antenna (Rx) in (b) is stronger than

the signal received in (a).

(Rx). The distance of P1 and P2 is measured and with (4.3), the signals’ strength arrived

at Rx through direct path, P1 and P2 were added together to obtain the signal strength

as detected at Rx. P1 and P2 are traced above and below the direct path. Using the same

method, an additional two rays are constructed at the two sides of the direct path

of the transmitted signal to the receiver. The addition of these five signals’ strength

results in the received signal strength as detected by the receiver antenna as shown in

Table 4.4. Also, predicted signal strengths using this hybrid method and using (4.3),

are compared with mean value of measured signal strengths from experiment. It is

assumed that 50% of the power is reflected in the calculation whenever a reflection

occurs. The results show that this hybrid method reduces the error margin.

Though this hybrid method is still in its infancy, the main point here is that wave prop-

agation characteristic (especially reflection) will have great impact on signal transmis-

sion and must be taken account in simulation planning.

Page 83


Table 4.4. Comparison of measured and predicted signal strength. D is the distance between

the transmitting antenna (with -20 dBm power delivered to the antenna) and receiving

antennas when both of them are facing each other; Pmea is the measured signal strength;

Plin and Phyb are the predicted signal strength using Eq. (4.3) and the hybrid method

respectively; ηlin is the difference between Pmea and Plin, and ηhyb is the difference

between Pmea and Phyb.

D (m) Pmea (dBm) Plin (dBm) ηlin (dB) Phyb (dBm) ηhyb (dB)

1 -39.62 -40.94 1.32 -40.33 0.71

2 -45.08 -48.47 3.39 -46.67 1.59

3 -47.14 -52.87 5.73 -50.21 3.07

4 -51.15 -55.99 4.84 -52.79 1.64

5 -51.38 -58.41 7.03 -54.86 3.48

6 -54.20 -60.39 6.19 -56.60 2.40

7 -57.07 -62.07 5.00 -58.10 1.03

8 -61.00 -63.52 2.52 -59.43 -1.57

35 -79.27 -86.31 7.04 -81.58 2.31

50 -82.27 -91.73 9.46 -86.98 4.71

As a side note, fluctuations of signal strength along the direct path (shortest path) be-

tween the transmission and the receiver antennas is observed. This is caused by con-

structive and destructive interference by reflected waves. Simulation models presented

in this research are not able to predict this and it is left as future research.

4.5.3 Radiation Pattern of Antenna

The power transfer between any two antennas is dependent on the angle of orientation

between them [60], unless they are isotropic radiators. A typical RFID antenna is a

6 dBi gain directional antenna and an example of the radiation pattern of a typical

Page 84


RFID antenna is shown in Fig. 4.6. This is the same radiation pattern presented in

Section 3.6.2, where it is used in the simulation software developed in Chapter 3. It is

reproduced here for easy referencing.

Despite the fact that in actual RFID deployment, it is very difficult to ensure all station-

ary RFID antennas are mounted in the fashion intended, with exact orientation, eleva-

tion and angles, the radiation pattern of any type of antenna to be deployed must be

acquired and fed into RFID simulation program as accurately as possible. Experience

shows that even two antennas of a same model manufactured by a same company will

have a slightly different antenna gain patterns. However the difference is small and

hence, the simulation program can assume the same gain pattern.

30

210

60

240

90

270

120

300

150

330

180 0

6 dBi 10

0

−10

−20

Figure 4.6. Polar plot of the antenna gain of a directional circularly polarised RFID antenna

with a gain of 6 dBi. (From Fig. 3.4.)

4.5.4 Simulation Result Interpretation and Analysis

A simulation program will use a model to compute the strength of signal at a certain

distance away from any transmitting antenna. There are many ways of presenting

a simulated result, either numerically or graphically. Fig. 4.7 shows an example of

simulated results in a graphical form, obtained using simulation software developed

in Section 3.7. In this example, two directional antennas with gain pattern shown in

Fig. 4.6 are located at position (50,50) and (400,50), both facing in the y direction and

Page 85

4.6 Second Carrier Sensing

Boundary

Figure 4.7. Boundary zone in simulation results. Results from simulation (vertical and horizontal

axes in units of 10 m, representing the y and x axes shown in Fig. 3.7, received power

in dBW).

with 1 W input power. Note that simulation results represent the received power by

isotropic antennas.

However, in real life applications, as shown in [60], the reception of a signal does not

exhibit a sharp cliff. For example, the line in the top of Fig. 4.7 shows signal strength

of -109 dBW. This line does not represent a clear threshold line between a region with

signal weaker than -109 dBW and region with signal stronger than -109 dBW. This is

due to uncontrollable factors that present themselves in a real life scenario, such as

fluctuation of path loss and antenna gain. However, a “boundary zone” (Fig. 4.7),

rather than a boundary line, can be specified in the simulation result to give good

estimate of the overall system performance before an actual deployment is carried out.

The size of the boundary zone should reflect the previously measured uncertainties in

the propagation loss which have already stated to be a few dB.


To maintain a low production cost, an RFID tag normally does not have the ability

to filter a valid interrogation signal in a channel from the other valid interrogation

Page 86


signals in adjacent channels. Hence the RFID tag can be “confused” when more than

one reader attempts to interrogate it. This problem is a type of collision problem and

is known as the Reader Induced Tag Confusion (RITC) problem (refer to Chapter 3 for

more discussion of collision problems). This situation is common in a dense reader

environment, such as a distribution warehouse.

Mitsugi first classified the “Listen Before Talk” (LBT) (refer Section 4.2.2) as first carrier

sensing and has suggested a second carrier sensing [71], to avoid the RFID tag’s being

“confused” by multiple interrogation signals, or in other words, to avoid the RITC

problem.

The flowchart of reader transmission which involves the use of second carrier sensing

in first carrier sensing enforced areas, is as illustrated in Fig. 4.8. Note that the first

carrier sensing part of this flowchart does not represent the whole picture of the first

carrier sensing (LBT) under the ETSI 302 208. For example, the case where a reader

antenna can choose to wait until the chosen channel is free for transmission after it

fails its first carrier sensing, is not shown in the flowchart. However, Fig. 4.8 does

represent a valid interpretation and a rational path of the LBT provision.

When a reader is ready to transmit, it chooses a channel and carries out LBT (or first

carrier sensing), to check whether there is any signal stronger than the regulated thresh-

old values in the channel of interest. If there is, it will choose a new channel and per-

form the LBT again. If there is no signal stronger than the threshold value, the reader

will carry out the second carrier sensing.

When the second carrier sensing is performed by a reader antenna, the reader antenna

is trying to use signals detected across the whole RFID band of operation as an approx-

imate estimate of whether the tags which this reader antenna is trying to interrogate,

will be “confused” by signals from other reader antennas or not. As mentioned before,

an RFID tag does not have the ability to differentiate signals of different frequencies.

Hence, to avoid this RITC problem, it is recommended that at a tag, the signal coming

from the intended reader antenna should be at least 15 dB higher than signals coming

from any other readers, to ensure a BER of at least 10−4.

Page 87


After a reader antenna of interest passes its first and second carrier sensing, it can start

to use its chosen channel to interrogate tags. If this reader antenna fails its second

carrier sensing, it has restart the entire process, which is to choose a new channel and

start the first carrier sensing again. It is hoped that after reselecting a new channel and

re-performing the first carrier sensing and by the time the new second carrier sensing

is carried out, the sources of interference (mostly mobile or hand held reader antennas)

would have been moved away from the nearby area of the reader antenna of interest.

The main difference between the first carrier sensing and the second carrier sensing is

that the first carrier sensing only checks for signal strength in the channel of interest

while the second carrier sensing checks for signal strength for the entire RFID band.

The threshold value for second carrier sensing is noticed, in the results to follow, to be

always higher than the first carrier sensing.

Reader ready to transmit

First Carrier Sensing

Choose a new channel

Pass?YesNo

Second Carrier Sensing

Lower

than

Threshold?

Yes

No

Reader Transmits

Figure 4.8. Second carrier sensing with “Listen Before Talk” provision. Flowchart showing the

stages before and after second carrier sensing is carried out. After a reader has found

a free channel, using first carrier sensing, in which to send a signal, this reader will

scan the entire frequency band allocated for RFID systems under the local regulations.

If this reader detects any signal in the frequency band higher than a pre-programmed

threshold value, it will decide not to transmit until successful second carrier sensing is

achieved.

Page 88


Mitsugi has presented results of second carrier sensing on random antenna placement.

However, we believe that antenna positioning is not random in nature. In practice,

when antennas are mounted, they follow a certain common pattern, such as side-by-

side or corner mounting, as shown in Fig. 4.9.

d

d 1

d

d 1

A

A

B

B

A

B

dd 1

(a)

(b)

(c)

Figure 4.9. Common antenna positioning. (a) face-to-face (b) side-by-side (c) corner.

It is obvious that second carrier sensing is not required on a repeating basis in an en-

vironment where only fixed position antennas are deployed, as in this case, they will

either fail or pass the second carrier sensing most of the time. For example, in a room

with two operating RFID readers with one transmitting antenna each, no matter which

channel one reader antenna has chosen to interrogate RFID tags, the signal strength

level within this reader antenna caused by the other reader will be detected and com-

pared with the pre-defined threshold of second carrier sensing. Since those antennas

are fixed position antennas, the signal strength will either be below or above the thresh-

old of second carrier sensing most of the time. However, in an environment where

hand held or mobile readers are used frequently, second carrier sensing can prevent

RITC.

The aim of the research presented in this section is to devise a sensible way to deter-

mine appropriate values to be used as the thresholds of the second carrier sensing for

Page 89


different tag positions and environments. The parameters of the reader positionings

for the investigation carried out are illustrated in Fig. 4.9.

Referring to Fig. 4.9, a tag is positioned at location X, at distance d directly in front of

antenna A. First, the signal from A (2 W ERP) received by the tag at X is calculated.

Then to have a BER of 10−4, a SNR (or CIR) of approximately 15 dB is needed. Hence,

the signal from B received by the tag at X must be at least 15 dB lower than the sig-

nal from A. From there, the minimum distance of antenna B from the tag for that d

is obtained. With this minimum distance, the minimum distance d1 between the two

reader antennas can be easily obtained. Using the path loss model shown in (4.2) with

a chosen environment factor n, the signal strength of the signal transmitted from an-

tenna B as detected by antenna A can be computed. This signal strength is used as the

threshold of second carrier sensing.

Hence when a tag is located distance d away from antenna A and antenna B is moved

closer than the just calculated d1, (Fig. 4.9 (a)), the second carrier sensing will prevent

antenna A from operation and will so avoid the RITC problem. If antenna B is at least

distance d1 away from antenna A, antenna B is allowed to transmit under the scheme

of second carrier sensing, as a BER of at least 10−4 has been ensured for the tag.

For cases (b) and (c) as shown in Fig. 4.9, the antenna gain pattern must be used to find

the optimum angle, θ, to have a suitable d1 and a suitable threshold value for second

carrier sensing.

4.6.1 Effect of the Position of Tags

It is highlighted in the previous section that the distance of the tag from its intended

reader antenna, d, is an important factor in deciding the suitable threshold value for

second carrier sensing. To observe the effect of the position of a tag, a simple path loss

model, (4.2), is used, instead of the suggested piece-wise linear path loss model for

RFID application, (4.3). The environment factor in (4.2), n, is set to 3.0 for all results

shown in this section.

Page 90


As the distance of the tag from antenna A increases, the minimum separation distance

d1 between antenna A and antenna B increases. Once d1 is obtained, the threshold

value corresponding to d1 is computed, the results being shown in Table 4.5. From the

results, it is concluded that the face-to-face orientation (Fig. 4.9 (a)) should be avoided,

as the minimum distance between the two antennas becomes relatively large, and this

creates a problem in a dense reader environment. A side-by-side orientation offers the

best results among these three orientations, with relatively low minimum separation

distance. This means more reader antennas can be positioned in an area to increase

RFID tag detection coverage.

Note that Table 4.5 does not imply whether a high threshold value is desirable or not.

The threshold values tabulated in that table simply correspond to the minimum sepa-

ration distances. In other words, to have a minimum d1, the chosen threshold values

for second carrier sensing cannot be lower than the values shown in the table. If not,

one reader antenna will be permanently shut off by the test of second carrier sensing.

4.6.2 Effect of Environment Factor

As discussed in Section 4.6.1, (4.2) is used for the study of second carrier sensing. To

observe the effect of environment factor, the tag distance from antenna A, d, is set to

3 m for all results shown in this section.

With the environment factor in (4.2), n, ranging from 2.5 to 4 with increments of 0.5, the

minimum d1 is calculated for each case together with the respective threshold value.

The results (shown in Table 4.6) are as expected. An incrementing n results in higher

path loss value, and hence a lower distance is required to attenuate the transmitted

signal to an acceptable level at the receiver side. Again, similarly to the analysis of the

effect of position of tags in Section 4.6.1, side-by-side orientation offers better flexibility

for antenna positioning, in that the interfering antenna may be positioned closer to the

antenna that is reading the tags.

Page 91


Table 4.5. Threshold values for second carrier sensing with respect to tag read range for

different antenna orientations. d is the distance between a tag and the nearest reader

antenna; d1 is the minimum distance between two reader antennas to have BER of at

least 10−4 for the tag, while the respective threshold values for second carrier sensing

are shown in the last column of the table.

d (m) d1 (m) Threshold (dBm)

Face-to-face

1 4.16 -8.58

2 8.32 -17.61

3 12.49 -22.89

4 16.65 -26.64

Side-by-side

1 1.30 -33.42

2 2.61 -42.50

3 3.91 -47.77

4 5.21 -51.51

Corner

1 4.28 -25.40

2 8.56 -34.43

3 12.84 -39.61

4 17.12 -43.46

4.6.3 Combining First and Second Carrier Sensing

If second carrier sensing is to be applied in a real life situation together with the first

carrier sensing, there remains a question on how to apply the positioning guidelines

already presented in this thesis in a correct sequence. The guidelines for first carrier

sensing (i.e. LBT) are discussed and presented in Section 3.9 while the guidelines for

second carrier sensing are presented in Section 4.6.1.

Page 92


Table 4.6. Threshold values for second carrier sensing with respect to environment factor,

n, for different antenna orientations. d1 is the minimum distance between two reader

antennas to have BER of at least 10−4 for the tag, while the respective threshold values

for second carrier sensing are shown in the last column of the table.

n d1 (m) Threshold (dBm)

Face-to-face

2.5 14.94 -19.36

3.0 12.49 -22.89

3.5 11.05 -26.52

4.0 10.11 -30.20

Side-by-side

2.5 4.28 -45.79

3.0 3.91 -47.77

3.5 3.64 -49.64

4.0 3.45 -51.51

Corner

2.5 15.24 -36.63

3.0 12.84 -39.61

3.5 11.45 -43.06

4.0 10.55 -46.63

Before going into an in-depth discussion, to avoid confusion, a few important points

are worth highlighting:

1. The antenna positioning guidelines for second carrier sensing depend on the

required tag read range, but the antenna positioning guidelines for first carrier

sensing do not.

2. The threshold of second carrier sensing was observed in our experiments to be

always higher than first carrier sensing. If this had not been the case, the first

carrier sensing would have been redundant. As the second carrier sensing checks

Page 93


for signal strength for the entire RFID band, which includes the channel chosen

for interrogation, if the threshold of second carrier sensing had been be lower

than first carrier sensing, it would have incorporated the function of first carrier

sensing with a stricter threshold.

3. As things currently stand, a failure in the test of first carrier sensing does not

imply a failure in the test of second carrier sensing, and vice versa.

Also in the beginning, two pieces of essential operating information are required, which

are the distance between a tag and the intended reader antenna, and the position and

orientation of reader antennas. The tag read range requirement depends on the ap-

plications of the RFID system, and may vary from 0.1 m up to 10 m. Some common

antenna orientations have been shown in Fig. 3.9.

Once the tag read range requirement, and the position and orientation of reader anten-

nas have been decided, there exist only two logical ways to proceed:

• The threshold value of the first carrier sensing is used as the positioning guideline

before that of the second carrier sensing.

Using the software developed in Section 3.7 with the technique in simulating an-

tennas operating in different channels in Section 3.9.4, antennas are ensured to be

able to operate (pick a free channel successfully) under first carrier sensing. Then,

the required tag read range is used to compute the minimum distance between

antennas, d1, to have at least BER of 10−4. If any two antennas are positioned

nearer than d1, they will be relocated at a further distance away (distance of at

least d1). The final step is to determine for each antenna the threshold value for

second carrier sensing that corresponds to d1, as these numbers will need to be

programmed into the readers connected to the antennas.

• The threshold value of the second carrier sensing is used as the positioning guide-

line before that of the first carrier sensing.

The tag read range is used to compute the minimum distance between antennas,

d1, to have at least BER of 10−4. If any two antennas are positioned nearer than

Page 94


d1, they will be relocated to a further distance apart. Then, using the software

developed in Section 3.7 with the technique in simulating antennas operating

in different channels in Section 3.9.4, antennas are ensured to be able to oper-

ate (pick a free channel successfully) under first carrier sensing. If any of the

reader antennas fails to obtain a free channel for transmission, antennas need to

be repositioned again. The final step is to determine the threshold values for

second carrier sensing that are required by each reader.

It can be seen that both ways are almost the same, with a difference in the sequence of

steps. Both ways determine the threshold value for second carrier sensing as the last

step, as this step is only carried out after the position of all the antennas involved has

been fixed and no longer requires any changes. In short, it is important to determine

the safe distance between reader antennas in the setting up of an RFID system with the

enforcement of first and second carrier sensing. After antennas have been placed in

the designed locations, the threshold values for second carrier sensing correspond to

the safe distance and intended tag read range are input into readers.

Once all the fixed position antennas are in place and pass the selected second carrier

sensing criteria, they should pass the second carrier sensing with these criteria unless

they are in a scenario where hand held or mobile reader antennas are present.

In the scenario where a hand held or mobile reader antenna enters an RFID enabled

environment with the enforcement of first and second carrier sensing, the readers will

readjust the channel distribution among antennas using first carrier sensing. However,

if ever a hand held or mobile reader antenna moves too close to an operating fixed

positioned antenna, that antenna will be switched off by second carrier sensing, to

avoid the RITC problem.

In a dense reader environment where antenna placement is very restrictive for both

first and second carrier sensing, fine-tuning methods for RFID systems are suggested

by [72], and are discussed in Section 5.6.

Page 95

4.7 Investigation of Specific Absorption Rate (SAR)


Specific Absorption Rate (SAR) is a measure of the rate at which radio frequency (RF)

energy is absorbed by the human body when exposed to radio frequency. This section

investigates the human radio frequency exposure level caused by 64 RFID reader an-

tennas positioned in the systematic array shown in Fig. 4.10. The focus is on the public

exposure (strictest level), and not the occupational exposure.

4.7.1 SAR Level for UHF RFID Operation

The safety levels discussed in this section are obtained from “Radiation Protection

Standard: Maximum Exposure Levels to Radiofrequency Fields - 3 kHz to 300 GHz,

Radiation Protection Series Publication No.3”, published by Australian Radiation Pro-

tection and Nuclear Safety Agency (ARPANSA) in 2002 [73].

The Australian RFID band spans from 920 MHz to 926 MHz, with a total of 12 channels.

In the case of public exposure, the maximum allowable power density, S, (in W/m2)

for frequency between 920 - 926 MHz, can be computed using the formula:

S =f

200(4.4)

where f is frequency in MHz.

From (4.4), a higher frequency will result in a higher maximum allowable power den-

sity. To have a conservative estimation, we chose a lower frequency to have a stricter

maximum allowable power density, which is 900 MHz. The maximum allowable

power density is then 4.5 W/m2.

4.7.2 Experiment Site

The 64 reader antenna array was used in an experiment to examine the interference

of a dense RFID reader environment to Australian GSM mobile telephone channels

Page 96


in Australia. This investigation of the SAR is to ensure human safety throughout the

duration of this experiment. The experiment site is shown in Fig. 4.11.

The main purpose of this study is to ensure the SAR 50 m away from the 64 anten-

nas array falls below the regulated maximum allowable power density. As shown in

Fig. 4.11(b), there is an office building 50 m in front of the test site. Also, it is the aim

of the study is to find out what is the safe distance from the 64 reader antennas array

when the maximum allowable power density is 4.5 W/m2.

4.7.3 Simulation of Power Density

As discussed in Section 4.7.2, the 64 antenna array is facing an office building. This

building is the location of interest and is 50 m away from the antenna site. Hence, 64

reader antennas were positioned as shown in Fig. 4.10 in the simulation software and

the simulated power density, Ssimu, is obtained at positions 50 m in front, 50 m on each

side and 50 m behind the 64 antennas.

The simulation of power density for the investigation on SAR is different from the

simulation of received power by a receiver antenna as presented in Section 3.7. This

is due to the fact that a human body is different from a reader antenna. It will absorb

RF power of different frequencies and the effect is cumulative. For example, a person

at a location exposed to 1 W/m2 contributed by a 900 MHz signal and to 1 W/m2

contributed by a 915 MHz signal, is in fact exposed to 2 W/m2 of power density.

This insight is consistent with the statement from [73] regarding the issue of simulta-

neous exposure to multiple frequency fields, which states that in general, exposure to

frequencies above 10 MHz are considered to be additive.

However, the software developed in Section 3.7 can be used to perform the required

calculation, as explained in the next section. In addition to this approach, a study of a

worst case scenario is presented in Section 4.7.5.

Page 97


50 m

50 m

50 m

Front Measuring Point

Back Measuring Point

SideMeasuringPoint

Figure 4.10. Arrangement of 64 reader antennas for SAR simulation. 64 antennas are arranged

in 4 lines, 16 antennas each. Adjacent lines are 2 m apart. Adjacent antennas in each

line are 2 m apart. All antennas are facing in the same direction (front).

4.7.4 Maximum Power for Every Channel

For the investigation of SAR, a simulation was carried out using the software presented

in Section 3.7. This method requires only a single run of simulation, with all the anten-

nas radiating at maximum power level at a chosen channel.

Page 98


(a) Side View (b) Rear View

Figure 4.11. The experiment site of the 64 reader antenna experiment. All the reader antennas

are facing the same direction. Note that antennas used in this experiment are from

many different manufacturers. Nonetheless, they are all directional patch antenna.

In actual fact, the 64 antennas are not operating in a same channel, but each antenna

will randomly pick one of the available 12 channels under Australian regulations.

However, as discussed before, the RF power absorption effect of a human being is

cumulative regardless of the frequency of RF signals. Hence, it does not matter an an-

tenna is operating in which channel. For simplicity in simulation, all the antennas can

safely be regarded to be radiating at a same channel.

Simulation results are shown in Table 4.7. The results indicate that at 50 m away in

four major directions, the power density is far below the allowable 4.5 W/m2.

Table 4.7. Simulation results on SAR investigation using maximum power for every channel.

Ssimu is the simulated power density in W/m2.

Position Ssimu

50 m in front 0.00585

50 m behind 1.00 × 10−5

50 m on each side 4.05 × 10−5

This simulation does not consider the power contained by side-band leakage as it is

believed that the contribution of side-band leakage is very small. The explanation

Page 99


to justify this point is based on the transmit masks presented in Section 3.9.3. The

dense-interrogator environments transmit mask (Fig. 3.5(b)) dictates that a side-band

leakage to the next immediate channel must be 30 dB below the signal strength of

the carrier signal. Hence, even when there exists a side-band leakage, its contribution

to the total radiation from an antenna will be more than 1000 times lower than the

contribution from that antenna within the channel in which it is radiating. For an area

with lower RFID reader density, the multiple-interrogator transmit mask (Fig. 3.5(a))

is used instead of the dense-interrogator transmits mask. The side-band leakage must

then be lower than 20 dB, and is still 100 times lower than a RFID signal operating in

that channel of interest.

Hence, even though the method of using maximum power slightly under-estimates

the power density, it provides a quick estimate of the power density at any location of

interest.

4.7.5 Worst-Case Scenario

The simulation results from Section 4.7.4 showed that it is safe at 50 m away from the

test site, with the method of using maximum power providing a slightly under but

quick estimation on power density. Hence it is in the interest of public safety that a

study of a worst-case scenario was carried out. Though physically unlikely or even

impossible, the following assumptions were made:

1. All antennas are represented by a single node and hence the resultant EIRP is

256 W EIRP (64 × 4 W EIRP).

2. Minimum path loss is used, which is the free space path loss.

3. There is a ground plane below resulting in perfect reflection.

4. The reflected wave is as strong as the direct wave and hence the signal is doubled

and the power density per unit area is quadrupled at the positions of interest.

Page 100


Table 4.8. Calculation results of the worst-case scenario.

Distance (m)With No Reflection Maximum Reflection

Power Density (W/m2) Power Density (W/m2)

1 20.37 81.48

2 5.092 20.37

3 2.263 9.053

4 1.273 5.092

5 0.815 3.259

6 0.566 2.263

7 0.416 1.663

8 0.318 1.273

9 0.251 1.006

10 0.204 0.815

15 0.091 0.362

20 0.051 0.204

25 0.033 0.130

30 0.023 0.061

35 0.017 0.067

40 0.013 0.051

45 0.010 0.040

50 0.008 0.033

55 0.007 0.027

60 0.006 0.023

Table 4.8 and Fig. 4.12 show the results of the worst-case scenario using the above

stated assumptions 1,2 and 3, and sometimes assumption 4. In the worst case, the safe

distance is 4 m away, but the public is not expected to enter this region. The public will

be in neighbouring buildings.

The locations of interest are located 50 m away from the 64 reader antenna array. The

power density level in the worst-case scenario at these locations is approximately 1135

Page 101


0.00

0.01

0.10

1.00

10.00

100.00

0 10 20 30 40 50 60

Distance (m)

Po

wer

Densit

y (W

/m2)

No Reflection

Max Reflection

4.5 W/m2

Figure 4.12. Plot of simulation results on SAR investigation. Plot of Table 1, highlighting

the safe distance in front of the 64 closely spaced reader antennas but with and

without reflection from a ground plane. In the worst case using the maximum reflection

assumption, the safe distance approximately 4 m.

of the allowable power density level in the frontal direction and much less in other

directions.

4.7.6 Recommendations on SAR

The recommendations on SAR issue of the 64 reader antenna deployment:

1. Simulation results show the power densities 50 m away in four major directions

are much lower than the maximum allowable power density.

2. The worst-case scenario shows that the safe distance in the most power intense

direction is 4 m. At 50 m away, The power density level in the worst-case scenario

at 50 m away is approximately 1135 of the allowable power density level.

3. The occupants below the roof would be shielded from the metallic roof or metallic

water proofing layer.

Hence it is safe to the public to carry out the 64 reader antennas experiment according

to the maximum exposure level as recommendation by ARPANSA.

Page 102


4.8 Conclusion

This research study has suggested ways of minimising the common errors that arise

from ignoring multi-path propagation, antenna orientation and pattern, and variation

of path loss as obstacles are traversed, in the simulation of RFID antenna deployments.

Some errors are unavoidable, but the extent estimated from empirical measurements

can be shown to be not severe. Hence it is essential not to interpret any simulation re-

sult without bearing all the possible errors and underlying assumptions in mind. This

study hopes to provide a clear view of the pitfalls for researchers interested in RFID

simulator development. Also, this study has highlighted EMC regulations which are

essential in understanding the actual implementation of an RFID system, in order to

produce a sensible simulator, which will definitely contribute to the vision of automat-

ing supply chains using RFID technology. Furthermore, the idea of second carrier sens-

ing was investigated to assist the deployment of RFID system. The SAR studies shows

that even with a large number of reader antennas located in a nearby area, and with

some penalising and unrealistic assumptions, the SAR exposure level is not hazardous

to the public at a relatively short distance away.

Page 103

Chapter 5

Reader Synchronisation

FOR a dense RFID reader deployment, such as in a warehouse,

where hundreds of readers will be positioned in a building, the

interference between all these readers must be studied carefully

to avoid the disruption of logistics operations. Strict RFID regulations and

standards have been imposed, trying to address the problem of reader col-

lision and also the problem of RFID devices interfering with other devices

operating in the same and nearby frequency bands. However, these guide-

lines and regulations are not entirely friendly for dense RFID reader deploy-

ment; in some cases it is not possible to have a feasible RFID system while

adhering to these regulations. Hence, this chapter proposes the synchroni-

sation of RFID readers to enable successful dense RFID reader deployment.

A case study targeted at European operations is presented in this chapter

to illustrate the actual synchronisation of RFID readers in real applications.

Some fine-tuning methods are also suggested to further improve the per-

formance of readers in a high reader density population area.

Page 105

5.1 Introduction

5.1 Introduction

Radio Frequency Identification (RFID) has received much attention recently as it is

widely believed that RFID can revolutionise supply chain management, complement-

ing barcodes as the main object tracking system. Several major supply chain operators

and retailers, such as Wal-Mart in the USA, have deployed RFID systems in some of

their supply chains [36]. Initial test runs of RFID deployment show encouraging re-

sults [74], and hence large scale RFID deployment is planned. However, before any

successful deployment can be achieved, some RFID issues have to be resolved. One of

them is the RFID reader collision problem, which is the focus of this chapter.

The term “reader collision(s)” is discussed extensively in [75] and [76]. In this chapter,

reader collision is simply defined as the phenomenon where an interrogation signal

from a certain reader disrupts the communication between a tag and another reader,

and this reader collision problem is potentially magnified in a dense reader environ-

ment, such as in a warehouse. Various regulatory and standardisation bodies have

tried to regulate the operations of RFID readers. In this research, the ETSI 302 208 as

introduced by the European regulatory body and the EPC Class 1 Generation 2 as rec-

ommended by EPCglobal are used as the basis of RFID reader operations. However,

as will be discussed in more detail in the latter part of this chapter, the restrictions that

are put on the operation of RFID readers are very strict, making it quite impossible to

have an uncoordinated large scale deployment of RFID readers. Hence, this chapter

further developes the idea of RFID reader synchronisation, to enable good RFID per-

formance in a dense reader environment, while adhering to strict regulations. The next

section introduces the ETSI 302 208 and EPC Class 1 Generation 2 Protocols and their

impact on RFID reader deployment. Section 5.3 explains the concept of RFID reader

synchronisation and how it adheres to strict regulations. Section 5.4 suggests possible

ways of implementing an RFID synchronisation system. A case study on RFID reader

synchronisation is presented in Section 5.5. Ways of fine-tuning RFID reader position-

ing are discussed in Section 5.6. Variations of possible reader synchronisation schemes

are presented in Section 5.7. Sections 5.8 offers views on the current progress of reader

Page 106

Chapter 5 Reader Synchronisation

synchronisation standardisation in Europe, followed by conclusion in Section 5.9. This

chapter is based extensively on [77] and [78], and is published as a book chapter in [79].

5.2 Background

5.2.1 ETSI 302 208

In the European Regulation as outlined in ETSI EN 302 208-1 V1.1.1 (2004-09) with “Lis-

ten Before Talk” (LBT) provision [48], a reader must “listen” and confirm that a partic-

ular channel is not occupied before it can use that particular channel to interrogate any

tag. The detailed discussion of LBT can be found in Section 3.6.4. The transmit power

and the corresponding threshold values are extracted from the above-mentioned ETSI

document and integrated into Table 3.2. They are reproduced as Table 5.1 here for

convenient reference.

Table 5.1. UHF RFID reader radiated power with corresponding threshold values for LBT.

Threshold value obtained from [48]. (From Table 3.2.)

Channel ERP (dBW) Threshold (dBW)

Up to 0.1 Up to -10 ≤-113

0.1 to 0.5 -10 to -3 ≤-120

0.5 to 2.0 -3 to 3 ≤-126

5.2.2 EPC Class 1 Generation 2 Protocol

“EPC Radio-frequency Identification Protocols Class 1 Generation 2 UHF RFID Pro-

tocol for Communication at 860 MHz - 960 MHz” [35], in short EPC C1G2, is the

standard protocol developed by EPCglobal for RFID devices for use within the sup-

ply chain. This protocol outlines the air interfaces and commands between an RFID

reader and an RFID tag. It also includes the spectrum management for RFID operation.

Page 107

5.2 Background

Frequency hopping or frequency agile systems are the suggested techniques. An allo-

cated frequency band, as allowed by local regulatory body, is divided into sub-bands

or channels. A reader will only use a certain channel for communication, not the en-

tire allocated frequency band. EPC C1G2 covers both dense reader mode and multiple

reader mode; multiple reader mode is for an environment where the number of simul-

taneously active readers is modest relative to the number of available channels while

dense reader mode is for an environment where the number of simultaneously active

readers is comparable to or more than the number of available channels. This docu-

ment only focuses on dense reader mode. In dense reader mode, for narrow bandwidth

(European 200 kHz) channels, it is suggested in this protocol that odd-numbered chan-

nels should be used for tag backscatter while even-numbered channels will be used for

reader interrogation. For a wide bandwidth channel (USA FCC 500 kHz channel [47]),

all available channels can be used for reader interrogation, as tag backscatters will be

located at the boundaries of these channels.

5.2.3 Problem in Dense Reader Environment

With the implementation of ETSI 302 208 and EPC C1G2, it is clear that when a reader

is operating at a certain sub-band or channel, this reader will effectively prevent other

readers from using that channel within an unacceptably large area. Section 3.9.3 has

presented detailed discussions and analysis on this matter and Table 5.2 (reproduced

from Table 3.5 for easy reference) summarises the minimum distance (calculated using

a piece-wise path loss model with variable environmental factor) between two anten-

nas connected to readers before one antenna operating at a certain channel will pre-

vent the other antenna from using that channel. It should be noted that these results

are obtained using a 0 dB isotropic receiving antenna, and do not represent any real

life situation, as a typical RFID antenna will be a directional antenna. Nonetheless, the

data presented in Table 5.2 gives sufficient evidence that a low threshold value for the

LBT as specified in ETSI 302 208 is severe enough to impede the reader deployment in

a dense RFID reader system.

Page 108


Table 5.2. Safe distance for different antenna configuration when the second antenna is an

isotropic radiator in accordance to LBT. (From Table 3.5.) Distances for channel

difference 1 obtained from Fig. 3.15(a) and Fig. 3.15(b). Results for other cases of

channel difference were obtained through separate simulations.

Channel DifferenceAntenna (Normal Configuration) Horizontal Antenna

Front (m) Side (m) Back (m) (m)

0 1400 350 210 320

1 180 45 30 50

2 130 25 15 35

3 95 20 10 30

5.3 Reader Synchronisation

Under the concept of reader synchronisation, all the RFID readers in a certain area,

for example all the readers in a warehouse, are networked together through a cen-

tral control unit. The connection method can be the common Ethernet connection, or

equivalent, and will be discussed in the next section. Since all the readers are linked

together, physically or wirelessly, they can be directed to execute commands at a same

time. Also, they can be assigned channels dynamically, so that the spectrum manage-

ment is optimised while the reader collision is minimised.

European regulation allows ten channels when maximum radiated power, 2 W ERP, is

used. Following the recommendation of EPC C1G2, under dense reader mode, five of

them, the even-numbered channels, are used for reader interrogation. All the readers

are “Listen Before Talk” compatible. They are configured to start to “Listen” at the

same time, and then at the end of the listen period, they can all synchronously start

to “Talk” as shown in Fig. 5.1. This is due to the fact that according to ETSI 302 208,

if there is no signal detected in the intended channel of interest, the “Listen” time is

fixed. Hence, all the readers, which start “Listening” at the same time, will start “Talk-

ing” at the same time. If a reader is turned on at a different time, or if a reader loses

Page 109

5.4 Actual Implementation

synchronisation that reader can be made to start again in synchronism with the rest of

the readers, after the last reader has finished its “Talk” session.

Reader 1

Reader 2

"Listen" "Talk"

"Listen"

Reader N "Listen"

"Talk"

"Talk"

Time

Figure 5.1. Synchronisation of all readers. All the readers start to “Listen” at a same time and

finish “Listen” at a same time too. (From Fig. 3.17.)


5.4.1 Connectivity

In actual implementation, the RFID readers must be able to communicate with each

other to enable synchronisation of the RFID readers. There are basically two ways to

connect all the readers; either using wired (physical) connection or using wireless con-

nection. A physically connected system or wired system cannot support mobile read-

ers. Also, a wired connection may suffer from data latency in the network. Ref. [80]

shows that time synchronisation in a wired network is possible, but will require ad-

ditional hardware and system reconfiguration. In the best case, the time difference

achievable can be better than 1 ms. In an actual implementation, synchronising sig-

nals may suffer latency in a busy network. However, since all the readers in a certain

vicinity can be linked as a local area network, the network congestion problem can be

reduced.

A wired system is often considered as a more reliable and a more secure communi-

cation method than a wireless communication. A wireless system signals through an

RF link. This link can use one of the five guard bands, mentioned in Section 5.2, for

sending a synchronising signal. A synchronising signal can be a signal with a special

Page 110


pattern. A wireless system can also use any existing wireless protocol such as Blue-

tooth technology. Wireless synchronisation supports mobile readers but is inevitably

vulnerable to interference (signal integrity problem), and unauthorised signal sniffing

(security problem). Both connectivity methods have their own advantages and disad-

vantages. The decision in choosing either of these two methods is largely dependent

on the positioning of the LBT sensor, which is discussed below.

5.4.2 Positioning of LBT Sensors

Reader

LBT Sensor

Reader Reader...

Antenna

Figure 5.2. Centralised LBT system. Readers in a nearby surrounding are connected to one LBT

sensor.

Reader

LBT Sensor

Reader Reader...

Antenna

Figure 5.3. Localised LBT system. Each reader has its own LBT sensor.

An LBT sensor of an RFID reader is responsible for detecting signals in the channel

of interest prior to transmission in that channel. This LBT sensor must have a power

sensitivity level better than -126 dBW as specified in [48]. If not, this LBT sensor will

not be able to function efficiently in determining whether there exists a signal with

Page 111


a power level higher than the power level specified in regulations in the channel of

interest. An LBT sensor can use the RFID antenna used for transmitting and receiving

signals in the communication with RFID tags. A LBT sensor can also use a separate

antenna connected to an RFID reader. Also, several RFID readers could share an LBT

sensor within a close vicinity. This is also known as a centralised system. A localised

system is where each and every RFID reader has its own LBT sensor. A centralised

LBT system is as shown in Fig. 5.2. The LBT sensor will constantly monitor all the

channels allocated for RFID operation, and dynamically assign available channels to all

the readers connected to it. The central control system has to be configured during the

initial setup of the system. A fine-tuned centralised LBT system offers high reliability.

However, it requires additional network configuration, to connect all the readers to the

LBT sensor using either wired or wireless connection. Also, a centralised LBT system

will not be able to be implemented effectively when mobile readers are dominant in

the surroundings. This is because although the readers can communicate with the

centralised LBT sensor through a wireless link, it is very difficult for the centralised

LBT sensor to estimate the position of mobile readers, and hence is not possible to

allocate the best channels for mobile readers.

For example, if two mobile readers operate simultaneously in an enclosed area, there

is a probability that the two readers move near to each other at some time. The cen-

tralised LBT sensor may at that time allocate very nearby channels to those two read-

ers and serious interference between those two readers may occur. Also, if two nearby

areas are running on different RFID wireless networks and they are un-coordinated,

interference with each other can occur, and in the worst case, cause a complete system

shut down. The coordination of wireless networks in different premises will be time

and cost consuming.

A localised LBT system is as shown in Fig. 5.3. Each reader has its own LBT sensor.

The LBT sensor can either use a separate antenna (Fig. 5.3), or use the same antenna

a reader uses to establish communication with an RFID tag within its interrogation

zone. As compared to centralised LBT system, a localised LBT system with wireless

connectivity enables new readers to be easily integrated into an existing system, with

Page 112


no additional cabling or setup needed. However, a localised LBT system has the prob-

lem of management of channel sharing, signal interference and possibly creation of

unwanted shielding. In actual fact, the connectivity of readers and the positioning of

an LBT sensor are closely related. In [81], a wired system and a centralised LBT are

linked together as one configuration, while a wireless system and a localised LBT are

linked together as another configuration.

5.4.3 Antenna Positioning

Truck Docks

h

AntennaA

AntennaBBest

ReadZone

Moderate Read Zone

Figure 5.4. Antenna positioning in dock door. A typical antenna setting at dock door, with h

being the height of the antenna from the base of a dock door.

The positioning of RFID interrogation antennas depends primarily on the application.

Detailed operational considerations for the deployment of an RFID system are pre-

sented in [63]. In this chapter, only one example will be given, which is the dock door

situation, as it will be used in the case study in the next section. A dock door is usually

2 to 3 m in width and approximately 3 m in height. The most effective way to create

an RFID interrogation zone is to position two antennas at the sides of the dock door,

face-to-face and with an height elevation, h, as shown in Fig. 5.4. The height elevation,

h, mainly depends on the average height of objects being shipped through the dock

Page 113

5.5 Case Study

door. A normal choice of h is between 0.5 to 1 m. Also, antenna A and antenna B

normally will be using different channels for tag interrogation. However, if antenna A

and antenna B are operating at the same time, and a tag is located in the middle of the

dock door, the tag may be “confused” by the interrogation signals from both of the an-

tennas with the result that the tag is misread. This effect is known as the tag confusion

problem. The discussion of this issue is outside the scope of this chapter but a simple

solution to this is to alternate the operation of antenna A and antenna B every query

cycle.

5.5 Case Study

A case study on dense RFID reader deployment at the dock doors of a warehouse is

presented here. As shown in Fig. 5.5, the green coloured rectangles represent trucks

loading or unloading goods at the dock doors of a warehouse. Each door is around 3

m in width, and has two RFID antennas facing each other for tag interrogation.

Since all the readers are synchronised in a way described in Section 5.3, they will

start “Listening” at a same time and will be assigned a channel for interrogation at

the end of “Listen” period. The assignment of channels will be geographically influ-

enced. Two readers assigned to be operating in the same channel will be as far apart

as possible. Also, the neighbouring antennas will be using channels as far apart as

possible. As illustrated in Fig. 5.6, the spectrum is split into ten channels, all five of

the odd-numbered channels are reserved for tag backscattering while all five of the

even-numbered channels are assigned for reader interrogation. Fig. 5.5 shows how the

channel assignment is done. The antenna on the furthest left is using channel 2 for

interrogation. The next antenna on its immediate right is using channel 8, which is six

channels away. Channel 10, though it is the furthest channel away, is not chosen. This

is because the arrangement of 2, 8, 4, 10, 6 gives best channel separation between every

channel.

Page 114


2 8 4 10 6 2 8 4 10 6 2 8Channel

Time"Listening"

Tru

ck

Tru

ck

Tru

ck

Tru

ck

Tru

ck

Tru

ck

Tru

ck

Tru

ck

Tru

ck

Tru

ck

Tru

ck

Tru

ck

Figure 5.5. Alternating of “Listening” and “Talking” mode. All the reader antennas at the

dock doors “Listen” at the same time and “Talk” together at some other time.

Channel

frequency

1 2 3 4 5 6 7 8 9 10

Figure 5.6. Channelling of the allocated frequency spectrum. Odd-numbered channels will be

used for tag backscatter signals while even-numbered channel will be used for reader

interrogation signals.

5.6 Synchronisation Fine-tuning

Fine-tuning of a synchronised RFID system, as presented in this section, can be carried

out to further reduce the tendency of reader collision. The fine-tuning methods dis-

cussed below include the reduction of output power, the reduction of overall reader

talking time, the use of external sensors, the use of RF opaque or absorbing materials,

and the frequent rearrangement of channels allocations.

Page 115


Figure 5.7. Reduction of output power to fine-tune reader synchronisation. Estimation of

required radiated power given that maximum read range corresponds to maximum ra-

diated power.

5.6.1 Reduction of Output Power

Although up to 2 W ERP can be used in a single or small population reader environ-

ment, in dense reader populations this higher power may not be necessary. Currently,

a state of the art reader can read up to 10 m. However, normal reading operations

do not require such a read range. In the case study presented in Section 5.5, the dock

doors of the warehouse are around 3 m in width. Since two antennas are positioned

facing each other in every dock door, the read range required is also around 1.5 to 2 m.

By reducing the radiated power of readers, the minimum distance between two anten-

nas using the same channel can also be reduced, this reduction being beneficial in a

dense reader environment. Fig. 5.7 gives an approximation to the reduction of output

power. In the far field region, using the Friis equation, the power received is the in-

verse function of the square of distance (r2). If the maximum read range corresponding

to maximum radiated power (2 W ERP) of a RFID reader is known, we can compute

the required radiated power for a shorter read range. For example, if the maximum

read range of a reader is 5 m using 2 W ERP (shown in Fig. 5.7), and if a read range of

only 2 m is required, the required radiated power can be lowered to 0.32 W ERP. This

estimation may not be accurate in real life due to complex electromagnetic propagation

phenomena, such as reflection caused by the surroundings objects, but it demonstrates

that power reduction is a viable option.

Page 116


5.6.2 Reduction of Overall Reader Talking Time

While it is possible to talk for 4 s, reader applications should be configured to talk

for only the time necessary to capture tag data. There is no optimum talking time. It

depends on the application and also the surroundings of the deployment zone. On-site

fine-tuning and measurements are needed before the reduction of talking time can be

carried out.

5.6.3 Use of External Sensors

Sensors can be used to turn RFID readers on only when tags are approaching thus fur-

ther reducing reader interference in that area. This will free up the channels allocated

for those antennas, and also to avoid unnecessary interference to other surrounding

reader antennas. For example, external sensors can be attached to the dock door in

the case study in Section 5.5. When the dock door is not in use, the designated RFID

readers would be switched off, as shown in Fig. 5.8. Optionally, the central control unit

can (as shown) dynamically shift the channels assigned for the antennas at door 3 to

door 4.

Channel

2 8 4 10 6 2 8 4 10 6

Tru

ck

Tru

ck

Tru

ck

Tru

ck

Tru

ck

Figure 5.8. Using sensors in an RFID system. Both the antennas at dock door 3 are switched

off when the absence of truck 3 is detected. This will free up the channels allocated for

those antennas, and also to avoid unnecessary interference to other surrounding reader

antennas.

Page 117


5.6.4 RF Opaque or RF Absorbing Materials

Another effective, but more expensive, way to reduce reader interference and collision,

is to utilise RF opaque or RF absorbing materials to contain the interrogating signal

within the designated zone of interrogation. For the case study presented in Section 5.5,

the use of such materials is shown in Fig. 5.9. Although there will still be some signal

leakage through the door openings, it would not have caused much interference. This

is due to the fact that the signal strengths at the sides of the antenna are relatively weak

as compared with the front of the antenna. According to [57], the gain at the side of

a typical RFID antenna is approximately 20 dB less than the gain at the front of the

antenna.

RF Absorbing Materials

2 8 4 10 6 2 8 4 10 6 2 8

Tru

ck

Tru

ck

Tru

ck

Tru

ck

Tru

ck

Tru

ck

Figure 5.9. Use of RF absorbing materials. All the antennas are separated using RF absorbing

materials. The antennas facing each other at the same door, are separated by at least

4 channels.

5.6.5 Frequent Rearrangement of Channels

Interrogating channels can be switched around every cycle of “Listen Before Talk”.

This is to prevent the jamming of the interrogation signal by any external noise. Fig. 5.10

shows a simple example on how the switching is done. There may be other more com-

plex switching methods involving higher artificial intelligence in the central control

unit, depending on the noise received from the surrounding environment, but these

await further development.

Page 118


Channel

8 2 6 10 4 8 2 6 10 4 8 2Channel

2 8 4 10 6 2 8 4 10 6 2 8

Figure 5.10. Channel switching within antennas. The channels arrangement are flipped from

left to right. Note that any face-to-face antennas are operating at least four channels

apart.

5.7 Variation of Synchronisation

In the previous sections, suggestions on the implementation of a real life RFID reader

synchronisation system are presented, together with some deployment options, such

as the connectivity of all the readers. Also, fine tuning methods are presented. In this

section, some of the interesting variations of RFID reader synchronisation schemes

are presented. These variations may not be readily incorporated into the suggested

methods mentioned in previous sections, but are presented here for future reference

and for completeness.

5.7.1 Separation of Transmitting and Receiving Channels

For the RFID, full power operation (2 W ERP) as governed by ETSI 302 208, only ten

channels are available, as shown in Fig. 5.6. However, as discussed in Section 5.2, there

are actually fifteen channels available for RFID in total. Five of the fifteen channels,

though used as guard bands, can be used for RFID operation with reduced maximum

allowable radiated power. There are three channels located lower in frequency than

the normal ten channels. These three channels can only be operated below 100 mW

ERP. There are two channels higher in frequency than the normal ten channels. These

Page 119

5.7 Variation of Synchronisation

1 2 3 4 5 6 7 8 9 10 1112131415

1 2 3 4 5 6 7 8 9 10

Complete

2 W

100 mW 500 mW

Figure 5.11. The complete frequency band allocated for RFID operation in Europe. Only

ten channels available for maximum allowable power operation (2 W ERP).

1 2 3 4 5 6 7 8 9 10 1112131415

1 2 3 4 5 6 7 8 9 10TxTx TxTx

Rx Rx Rx Rx Rx

Figure 5.12. Separation of transmitting and receiving channels in the frequency band al-

located for RFID operation in Europe. All the Transmitting (Tx) channels are

separated by two Receiving (Rx) channels as compared to the complete frequency

band shown in Fig. 5.11.

two channels can be operated below 500 mW ERP. The complete frequency range for

RFID operation, with respective regulated power level is as shown in Fig. 5.11.

The channel numbering system shown in Fig. 5.6 is included in Fig. 5.11, along with

a new channel numbering system to simplify the discussion hereon. Channel 4, 7, 10

and 13 are assigned to be the reader transmitting channels while the tag reply channels

are the four channels beside the transmitting channels [82]. For example, transmitting

channel 4 uses channel 2, 3, 5, and 6 for tag reply. Although the transmitting channels

are reduced from a total of five down to four, the transmitting channels are placed

two channels away rather than one channel away from the neighbouring channels.

(Fig. 5.12). From the transmit mask shown in Fig. 3.5, an improvement of 5 dB can

be obtained. Hence with the reduction of interference between transmitting channels,

readers can be placed nearer to each other.

Page 120


5.7.2 Separation of RFID and Non-RFID Signals

Another variation of synchronisation is to differentiate an RFID signal from a non-

RFID signal. A method using signal recognition is presented in [83]. The idea is that

all the RFID readers in a certain region could possibly be treated as a single entity

in the regulations as outlined in the ETSI 302 208. Hence it is only required in this

interpretation that signal interference between all the RFID readers and the rest of the

short-range devices be avoided. If this concept is valid, the interrogation signals of

RFID readers are not treated as a signal in a channel when an LBT test is carried out.

The main advantage of this method is that a lot of readers can be deployed in a small

confinement area. However, reader antenna positioning can become more challenging,

as all the readers can choose any channel for transmission as long as there is no other

type of short-range device around.

5.8 Updated Progress on Development of RFID Reader

Synchronisation

The regulation of RFID reader synchronisation in the European countries is governed

by ETSI. TG34: Radio Frequency Identification Devices (RFID), a technical work group

within ETSI has been field-testing late in 2006 in an operational distribution center us-

ing a synchronised RFID system [84]. The testing involved up to 36 adjacent portals

operating simultaneously. Using identical pallets comprising 63 “RFID unfriendly”

cartons, a read rate of better than 98.5 % is recorded. In a near future, RFID reader

synchronisation would be included in ETSI TR 102 436 “Electromagnetic compatibil-

ity and Radio spectrum Matters (ERM); Improved spectrum efficiency for RFID in the

UHF Band”, and be harmonised with the current EN 302 208. Also, with the standard-

isation of RFID reader synchronisation, the mandatory use of “Listen Before Talk”,

which is a deterrent to large-scale deployment of RFID system, will be lifted, allowing

a better-performance and better coverage RFID systems in Europe.

Page 121

5.9 Conclusion

5.9 Conclusion

This chapter has identified synchronisation of RFID readers as a mechanism to assist in

RFID reader deployment in dense reader environments. Some implementation meth-

ods, and several fine-tuning methods are also presented in optimising the performance

of a synchronised RFID system. As compared with conventional unsynchronised RFID

systems, a synchronised RFID system can offer more coverage, less reader collision or

interference, while strictly following the European regulations and the EPC C1G2 rec-

ommendation, and can, with variation of the normal operating procedure, deal also

with the effects of tag confusion. However, these benefits require the use of more

complex hardware and hence can marginally increase deployment costs. Reader syn-

chronisation has not been tested in a real situation, and hence will require future study

in this area.

Page 122

Chapter 6

RFID Tag Antenna Design

THIS chapter focuses on the design of RFID tag antennas, begin-

ning with the introduction of relevant antenna theory. The chal-

lenges in designing small RFID tag antennas are discussed, fol-

lowed by the presentation of the methodologies and tools used in the pro-

cess of simulation and fabrication of antenna prototypes. The design and

analysis of various types of HF and UHF RFID tag antennas, mostly indus-

trial driven, are presented at the end of this chapter.

Page 123

6.1 Introduction

6.1 Introduction

The design of an antenna for an RFID tag depends heavily on the application. In the

current RFID worldwide market there is a wide variety of RFID tags with very different

antennas of different sizes, shapes, materials, and operating frequencies. This is to

cater for the different needs of various applications.

Often, an RFID tag antenna designer will be given the specifications of which fre-

quency band the tag is to be operate in, the location on the product where the tag

is going to be attached, and the read range performance requirement for the tag.

There are standard design procedures for an RFID tag antenna. They are very useful

as a reference when the RFID tag antenna is to be attached to an object that does not

have a space constraint and which is to be deployed in an RFID friendly environment.

However, there are cases where the design of an RFID tag antenna is challenging.

Throughout the author’s research, some very challenging RFID tag antenna design

problems have been encountered, mostly presented by industry partners.

This chapter will contain discussion in Section 6.2 of the antenna theory involved in

designing an RFID tag, definition in Section 6.3 of the antenna parameters used in any

discussion throughout this thesis, and presentation in Section 6.4 of equivalent circuits

of an RFID antenna. With a firm grasp of the fundamental electromagnetic properties

of antennas, the discussion is extended in Section 6.5 to the limitations and challenges

of designing an RFID tag antenna.

The preparation of HF and UHF RFID tag chips used in the experiments is discussed

in Section 6.6 while the range of HF and UHF RFID readers and RFID reader antennas

are documented in Section 6.7.

Section 6.8 explains the fabrication process of designed tags presented in this thesis.

Also, the simulation and prototyping software is included as Section 6.9. The fabri-

cation process and testing methods of a prototype antenna are also discussed in Sec-

tion 6.9. A further investigation of antenna performance measurements can be found

in Chapter 8.

Page 124

Chapter 6 RFID Tag Antenna Design

Section 6.10 begins with the design of simple HF RFID tag antennas, and then moves to

a consideration of more complex HF RFID tag antennas which cater for specific needs

and specifications.

Section 6.11 is very similar to Section 6.10, except the focus is on the UHF RFID tag

antennas rather than the HF RFID tag antennas.

6.2 Antenna Theory

A major focus of this thesis is on the design of HF and UHF RFID tag antennas. Inten-

sive background studies based on [85] and [86] were carried out to establish a strong

foundation in antenna theory and to distinguish the difference between the design of

an HF RFID tag antenna and the design of a UHF RFID tag antenna. This section lists

several important points of the related antenna theory.

• Common terms in electromagnetism

The common terms in the study of electromagnetism are defined by The Interna-

tional System of Units [3] and are shown in Table 6.1.

Table 6.1. SI units for common terms used in antenna theory.

Electric field strength E Vm−1

Electric flux density D Cm−2

Magnetic field strength H Am−1

Magnetic flux density B Wbm−2

Permittivity ε Fm−1

Permeability µ Hm−1

• Maxwell’s Equations

The fundamental antenna theory is based on Maxwell’s equations, first intro-

duced by James Clerk Maxwell in unifying the theories of electromagnetism. The

Page 125

6.2 Antenna Theory

four Maxwell’s equations are the combination of Faraday’s Law, Ampere’s Law

as modified by Maxwell, Gauss’s Law for electric flux and Gauss’s Law for mag-

netic flux.

In differential form:

∇× E = −∂B∂t

(6.1)

∇× H = J +∂D∂t

(6.2)

∇ · D = ρ (6.3)

∇ · B = 0 (6.4)

• Sources and vortices

To visualise the electric, magnetic or electromagnetic fields, the concepts of a

source and a vortex are introduced. A source like field is as shown in Fig. 6.1(a)

and vortex like field is as shown in Fig. 6.1(b). Normally, an electric field is con-

sidered as a source like field while a magnetic field is considered as a vortex like

field.

(a) Source (b) Vortex

Figure 6.1. The difference between source and vortex.

• Boundary conditions

A brief treatment of electromagnetic boundary conditions is presented here, with

its important impact on RFID tag antenna design highlighted. A complete treat-

ment of boundary conditions can be found in [87].

Electromagnetic waves will undergo changes when they traverse from one medium

into another medium. If one of the media is a conductor, such as a metallic sur-

face, there can only exist a tangential magnetic fields and a normal electric fields

Page 126


at the boundary. Fig. 6.2 shows the changes on electric and magnetic field near a

metallic surface.

Charge

Metallic Surface

Wire

Figure 6.2. Boundary conditions affecting electric and time varying magnetic field. If the

metallic surface is a perfect conductor, there are only tangential magnetic fields and

normal electric fields at the boundary.

• Spherical Coordinate System

In the study of antenna theory, a spherical coordinate system is often used over

the conventional cartesian coordinate system. This is due to the nature of the field

pattern when it can be described in the simplest manner using a spherical coordi-

nate system. Fig. 6.3 shows the configuration of a standard spherical coordinate

system.

P( , , )r è

z

y

x

è

r

Figure 6.3. Spherical coordinate system. Any point in space can be specified by using (r,θ,φ)

coordinates shown above.

• Near field and far field

The analysis in this section is from [85] and [88] . The analysis applies to the case

where r > 0.

Page 127

6.2 Antenna Theory

The E and the H fields for an infinitesimal electric dipole can be expressed in

spherical coordinates as:

Hr = Hθ = 0 (6.5)

Hφ =β2 I0l4π

[j

(βr)+

1(βr)2

]e−jβrsinθ (6.6)

Er = ηβ2 I0l4π

[2

(βr)2 − 2j(βr)3

]e−jβrcosθ (6.7)

Eθ = ηβ2 I0l4π

[j

(βr)+

1(βr)2 − j

(βr)3


Eφ = 0 (6.9)

where the subscripts r, θ and φ, indicate the direction of the fields according to

spherical coordinate system; η is the wave impedance which equals 377 Ω in free

space; β is the propagation constant and is computed using λ2π ; I0 is the current

in the electric dipole; l is the length of the electric dipole; r is the distance from a

fixed reference point. It can be easily proven from (6.5) - (6.9), and by using the

formula for time-average power density:

Wav =12

Re[E × H∗] (6.10)

When r = λ2π , the first and the third terms in the bracket of (6.8) are the same in

magnitude but opposite in sign, and hence will cancel out each other. However,

when r < λ2π , the magnitude of the third term in the bracket of (6.8) will be greater

than the second term in the same bracket. In the case of r λ2π , the third term

in the bracket of (6.8) will completely dominate the other two terms. When this

happens, the energy storage field will be the dominant part while the power flow

(radiating field) will be negligible.

On the other hand, at a distance r > λ2π , the first terms of (6.6) and (6.8) will be

dominant. In this zone, power flow (radiating field) will be the dominant part

while the energy storage field will be negligible in comparison.

A similar inspection can be carried for a magnetic dipole. The E and the H fields

of a small loop or infinitesimal magnetic dipole can be expressed as:

Page 128


Er = Eθ = Hφ = 0 (6.11)

Eφ =β2 jωµ0 IA

4π

[j

βr+

1(βr)2


Hr =1η

β2 jωµ0 IA4π

[2

(βr)2 − 2j(βr)3

]e−jβrcosθ (6.13)

Hθ =1η

β2 jωµ0 IA4π

[j

βr+

1(βr)2 − j

(βr)3


where the subscripts r, θ and φ, indicates the direction of the fields according to

spherical coordinate system; η is the wave impedance which equals 377 Ω in free

space; β is the propagation constant and is computed using λ2π ; I is the current in

the magnetic dipole; A is the area enclosed by the magnetic dipole (loop); and r

is the distance from an origin at the centre of the loop.

When r λ2π , the third term in the bracket of (6.14) and the second term in the

bracket of (6.13) will completely dominate the other two terms (energy storage

field dominant), while at a distance r > λ2π , the first terms of (6.12) and (6.14) will

be dominant (radiating field dominant).

Hence, r = λ2π is defined as the boundary of the near field and far field. The near

field is the energy storing field while the far field is the energy radiating field. For

the operation of HF and UHF, this boundary is at a distance of 3.5 m and 0.052 m

respectively.

• Coupling and radiation

Coupling and radiation are two different means of power transfer from a point to

another point without using any physical connection between those two points.

In other words, power is transferred using a wireless connection; its possibility is

consistent with the Maxwell’s equations.

The coupling that normally takes place in the near field, which has been defined

previously, is in the form of either inductive or capacitive coupling. In such cou-

pling, energy flows out of the source in the form of electric or magnetic field and

flows back to the source. Ideally, there will be no energy loss if there is no resis-

tive load near the field. Hence, a near field is also known as the energy storage

Page 129

6.2 Antenna Theory

field. If there is a load in the near field, energy will be transferred from the source

to the load through coupling.

Radiation normally takes place in the far field, which has been defined previ-

ously. In the case of radiation, energy propagates or radiates away from the

source and never returns to the source. When a load adsorbs radiated energy,

the source will not be significantly affected.

However, the term “couple” or “coupling” is also commonly used in the litera-

ture with a different meaning. For example, the phrase “antenna A is coupled to

the magnetic field” does not mean that the antenna A is coupled in the near field.

It means that antenna A is sensitive to the magnetic field, whether it is in the near

or far field.

• Power Transfer

In the near field operation, to estimate the power received by an RFID tag chip,

coupling volume theory as introduced in [89] is applied.

Coupling volume is a figure of merit. For a magnetic dipole, the equations in-

volved are given by:

Vc =

Reactive power flowing in the untuned

label coil when it is short circuited

Volume density of reactive power

created by the interrogator at the label position

(6.15)

Vd =

Reactive power flowing in the inductor

of the interrogator field creation coil

Volume density of reactive power created

by the interrogator at the label position

(6.16)

If there are two magnetic dipoles resonating with quality factors Q1 and Q2 re-

spectively, the ratio between the power dissipated in the second magnetic dipole,

P2, and the power dissipated in the first magnetic dipole, P1, is given as:

Page 130


P2

P1=

Vc

VdQ1Q2 (6.17)

If the first magnetic dipole is an RFID reader antenna (transmitting antenna with

quality factor Q1), (6.17) can estimate the power received by an RFID tag, where

the RFID tag antenna is acting as the second magnetic dipole with quality factor

Q2.

The coupling volume for an electric dipole is slightly different in definition as

compared to the coupling volume for a magnetic dipole:

Vc =

Reactive power flowing in the untuned label capacitor

when it is open circuited

Volume density of reactive power

created by the interrogator at the label position

(6.18)

Vd =

Reactive power flowing in the capacitance

of the interrogator field creation electrodes

Volume density of reactive power created

by the interrogator at the label position

(6.19)

Similarly to the case of two magnetic dipoles, if there are two electric dipoles res-

onating with quality factors Q1 and Q2 respectively, the ratio between the power

dissipated in the second electric dipole, P2, and the power dissipated in the first

electric dipole, P1, is given as:

P2

P1=

Vc

VdQ1Q2 (6.20)

However, (6.17) and (6.20) are only meant for near field power transfer. In the far

field zone, radiation theory is often applied.

From radiation theory, the power transfer from the transmitting antenna to a loss-

less receiving antenna can be calculated using:

Page 131

6.2 Antenna Theory

Pr

Pt= gtgr

(λ

4πd

)2

(6.21)

where λ = wavelength; Pt = transmitted power; Pr = available received power;

gt = transmitter antenna gain; gr = receiver antenna gain; d = separation distance

between antennas.

It must be noted that (6.21) does not take in consideration of any mismatch be-

tween the antenna and the load. An RFID reader antenna often has a good match

with an RFID reader RF front end. However, owing to size constraints, small

RFID tags may not have a match between the RFID tag antenna and the RFID

tag chip. This mismatch loss must be taken into account using (6.22), and will be

discussed next.

• Maximum power transfer

A passive RFID tag normally does not have its own power source and it is pow-

ered by an interrogating antenna using CW transmission.

Hence to enable maximum read range, power transmission between the tag an-

tenna and the tag chip must be maximised. It can be proven easily that the maxi-

mum power transfer occurs when the impedance of the antenna is the conjugate

of the impedance of the tag chip. Note that the tag antenna is not a lossless an-

tenna.

If there is no conjugate match, the power loss can be computed through [85]:

PlostPavailable

=∣∣∣∣Zant − Z∗

cctZant + Zcct

∣∣∣∣2

(6.22)

where Zant is the input impedance of the antenna (in our case, Zant refers to the

designed RFID tag antenna) and Zcct is the input impedance of the circuit which

is connected to the antenna at its input terminals (in our case, this is the input

impedance of an RFID tag chip used in prototyping).

• Polarisation

Page 132


Polarisation is the curve traced by the end point of a vector representing the in-

stantaneous electric field [85]. In literature, there are three types of polarisation,

linear, circular and elliptical, though in fact a linear or a circular polarisation is a

special case of elliptical polarisation.

An RFID tag antenna is usually linearly polarised. However, the orientation of

an RFID tag antenna may change rapidly as the object it is attached to is being

transported along the supply chain. If a linearly polarised RFID reader antenna

is deployed and if the RFID tag antenna is aligned orthogonally to the reader

antenna, the reader will not be able to communicate with the RFID tag. Based on

this, RFID reader antennas are often circularly polarised.

Polarisation efficiency is defined as “the ratio of the power received by an an-

tenna from a given plane wave of arbitrary polarisation to the power that would

be received by the same antenna from a plane wave of the same power flux den-

sity and direction of propagation, whose state of polarisation has been adjusted

for a maximum received power” [85].

It can be observed easily, without mathematical proof, from a Poincare sphere

[86], that the polarisation efficiency between a circularly polarised RFID reader

antenna and a linearly polarised RFID tag antenna is 0.5 (or -3 dB).

• Scattering Parameters

Scattering parameters are mostly used in the study of distributed circuits involv-

ing transmission lines. In general, a scattering parameter represents the ratio

between an incident wave, a, to a port of interest, and an emerging wave, b, from

the same port or from another port. In particular,

Sij =bi

aj(6.23)

where aj is a known signal entering through port j; bi is the signal detected in

port i when port i is fed from a matched source or terminated with a matched de-

tector. It should be noted that all other ports (if there are any) must be terminated

with a matched load. The a and b parameters have the property that they are

Page 133

6.3 Antenna Parameters

in phase with the voltages (or transferred electric fields) of the travelling waves

and the amplitudes and units are adjusted so that the square of the magnitude

represents the power flow. In the usual case, the characteristic impedances, Z0,

of any transmission line involved are assumed to be of the same value, and the

scattering parameters can be defined as ratios of transmission line voltages.

In the measurement of the performance of a designed antenna, the antenna is

connected to a network analyser using a coaxial cable (a transmission line). A

normal antenna can be considered as a single port network. Hence, the most

common S parameter is S11. S11 is the ratio of the reflected signal to the incident

signal, also known as the input reflection coefficient, Γ.

Alternatively, input reflection coefficient can be computed using:

Γ =ZL − Z0

ZL + Z0(6.24)

where ZL is the input impedance of the antenna under test, and Z0 is the charac-

teristic impedance of the coaxial cable connected to the antenna. Z0 is normally

purely resistive and hence (6.24) is always in magnitude less than 1.

S21 is sometimes used when two antennas are involved (one as the transmitter

and the other one as the receiver). S21 describe the transmission quality between

those two antennas. In the case of two passive antennas, S21 = S12, due to reci-

procity.

In measuring an antenna using a network analyser, the reflection measurement

is defined as the ratio of reflected power to incident power, which is |S11|2.


Common terms to describe the characteristics of an antenna, also known as the antenna

parameters, are referred to in various publications [85, 86, 90, 91, 92]. Definitions are

obtained from the “IEEE Standards Definitions of Terms for Antennas” [90] to avoid

any ambiguity in this thesis. Note that an antenna is assumed by Section 1.1 of [90] (not

Page 134


the Section 1.1 of this thesis) to be linearly polarised. This is not the case in RFID ap-

plications as an RFID reader antenna is usually circularly polarised. Special notes will

be added for any parameter which requires extra explanation for the case of circularly

polarised antenna.

1. Directivity

Directivity of an antenna is the ratio of the radiation intensity in a given direc-

tion from the antenna to the radiation intensity averaged over all directions. The

average radiation intensity is equal to the total power radiated by the antenna di-

vided by 4π. If the direction is not specified, the direction of maximum radiation

intensity is implied.

2. Gain

In simple form, the gain of an antenna is obtained through the reduction the di-

rectivity of an antenna by the amount equal to the dissipative loss of the antenna.

If an antenna does not have dissipative loss, the gain of this antenna is equivalent

to the directivity of this antenna.

Gain is defined as the ratio of the radiation intensity, in a given direction, to the

radiation intensity that would be obtained if the power accepted by the antenna

were radiated isotropically. Gain does not include losses arising from impedance

and polarisation mismatch.

Special notes: The gain of a linearly polarised antenna is quoted in “dBi”. In

the case of a circularly polarised antenna, antenna manufacturers will sometimes

quote the gain of an antenna using the term “dBic”, though some manufacturers

still prefer to use “dBi”. This may create some confusion. A simple rule is that

“dBic” of a circularly polarised antenna is 3 dB higher than an equivalent “dBi”

circularly polarised antenna.

3. Radiation Pattern

The radiation pattern is the spatial distribution of a quantity that characterises

the electromagnetic field generated by an antenna. The quantities that are most

Page 135


often used to characterise the radiation from an antenna are proportional to, or

equal to, power flux density, radiation intensity, directivity, phase, polarisation,

and field strength. When the quantity is not specified, an amplitude or power

pattern is implied.

In this thesis, the radiation pattern plots show either the gain or directivity of the

antenna under test, unless otherwise stated.

4. Polarisation

See Section 6.2.

5. Input Impedance

The impedance presented by an antenna at its terminals. Impedance of an an-

tenna, Z, is normally presented in the form of R + jX, where R is the resistance

of the antenna and X is the reactance of the antenna.

6. Radiation Efficiency

The ratio of the total power radiated by an antenna to the net power accepted by

the antenna from the connected transmitter.

7. Quality Factor

Quality factor, Q, of a resonant antenna is the ratio of 2π times the energy stored

in the fields excited by the antenna to the energy radiated and dissipated per cy-

cle. For an electrically small antenna, it is numerically equal to one-half the mag-

nitude of the ratio of the incremental change in impedance to the corresponding

incremental change in frequency at resonance, divided by the ratio of the antenna

resistance to the resonant frequency.

8. Bandwidth

The range of frequencies within which the performance of the antenna, with re-

spect to some characteristic, conforms to a specified standard.

9. Radiation Resistance

Page 136


Radiation resistance is the ratio of the power radiated by an antenna to the square

of the r.m.s. antenna current referred to a specified point. The total power radi-

ated is equal to the power accepted by the antenna minus the power dissipated

in the antenna.

6.4 Resonant Circuit for Antenna

To estimate the performance of antenna, it is useful to visualise an antenna as its reso-

nant circuit, which consists of just resistors, inductors and capacitors. From a resonant

circuit, it is convenient to compute the quality factor of an antenna. However, it must

be stressed that this is purely an estimation as at high frequency, a distributed circuit

is more accurate than a discrete component resonant circuit. Nonetheless, a discrete

circuit is sufficient to provide a good initial estimation of the performance of a new

design.

There are three types of simple resonant circuits; series resonant circuits, parallel res-

onant circuits, and practical resonance circuit. The details of for these three resonant

circuits can be found in [93]. Some important results are shown here for the discussion

on how an RFID tag antenna or RFID tag chip can be represented by these equivalent

circuits.

• Series Resonant Circuit

A series resonant circuit consists of a resistor, Rs, an inductor, Ls, and a capacitor,

Cs, linked together in series as shown in Fig. 6.4.

R L

CV s

ss

Figure 6.4. Series resonant circuit.

The resonant frequency of this series resonant circuit is given by:

Page 137

6.4 Resonant Circuit for Antenna

f0 =1

2π√

LsCs(6.25)

The quality factor of this series resonant circuit is given by:

Q =XL

Rs(6.26)

where XL the reactance of the inductor at resonance.

The bandwidth of this series resonant circuit is given by:

BW =f0

Q(6.27)

where f0 is the resonance frequency of the circuit.

• Parallel Resonant Circuit

A parallel resonant circuit consists of a resistor, Rp, an inductor, Lp, and a capaci-

tor, Cp, aligned in parallel as shown in Fig. 6.5.

R L CI ppp

Figure 6.5. Parallel resonant circuit.

The resonant frequency of this parallel resonant circuit is given by:

f0 =1

2π√

LpCp(6.28)

The quality factor of this parallel resonant circuit is given by:

Q =Rp

XL(6.29)

The bandwidth of of this parallel resonant circuit is given by:

Page 138


BW =f0

Q(6.30)

• Practical Parallel Resonant Circuit

The parallel resonant circuit shown in Fig. 6.5 is often known as the ideal parallel

resonant circuit. In the case of antenna design, no actual resistor will be added in

the antenna. The resistor in a resonant circuit of an antenna normally represents

the losses introduced by the inductor. Hence, as an alternative, the circuit of

Fig. 6.6, also known a practical parallel resonant circuit, may be investigated.

r

L

CI p

p

Figure 6.6. Practical parallel resonant circuit.

Though Fig. 6.6 better characterises an antenna, it is easier to analyse mathe-

matically when an antenna is represented in the form of ideal parallel resonant

circuit. It can be done be transforming, approximately, the losses in inductor, r, to

the parallel resistor, Rp using:

Rp =(ω0Lp)2

r(6.31)

where ω0 is the angular resonant frequency of the circuit.

Different types of RFID tag chips are discussed in Section 6.6. For the case of HF chip,

the input impedance is just the input capacitance. Hence, the capacitor (Cs in Fig. 6.4

and Cp in Fig. 6.5) may be the input capacitance of the HF chip. The inductor and the

resistor represent the inductance of the HF coil antenna and the losses within the coil

respectively.

Page 139

6.5 Challenges in RFID Tag Antenna Design

The case of UHF RFID tag chips is more complex, in the sense that the input impedance

of a UHF RFID tag chip is represented by a resistor in parallel with a capacitor. For

example, a Texas Instruments RI-UHF-STRAP-08 UHF Gen2 Strap [94] (More details

in Section 6.6) has an input impedance specified by 380 Ω‖2.8 pF, or 9.7866 − j60.1926

in series.

A UHF RFID tag antenna normally has a resultant inductive input impedance to min-

imise the power transfer loss dictated by (6.22) between the antenna and its UHF RFID

tag chip. Hence, it can be represented by an inductor and resistor. Note that, this does

not mean an antenna does not have self-capacitance. It is just that at the frequency of

interest, the antenna is inductive in impedance.

A resonant circuit can then be formed using the inductor and resistor from the antenna,

together with the capacitor and resistor from the chip.

Note also that the bandwidth of a UHF RFID tag does not equal to the bandwidth of a

UHF antenna. This is caused by the additional resistance from the input impedance of

the UHF chip.


The design procedures for an HF RFID tag antenna are mature and well-documented,

such as in [95]. Basically, an HF RFID tag antenna provides enough inductance to

resonate with the HF RFID tag chip input capacitance, to operate at the frequency of

13.56 MHz. The HF tag is designed to have a reasonably high quality factor so as to

have a satisfactory read range (Some HF tags are purposely designed to have a low Q

as specified in 18000 Part 3 Mode 2). The read range of an HF system is often limited

by the reader antenna size, not the HF tag antenna, though in some cases the limiting

factor is the size of the tag antenna itself.

However it is not the same case for the design of UHF RFID tag antennas. Accord-

ing to [96], the main performance limitations of passive UHF RFID systems are chip

sensitivity, antenna gain, polarisation and impedance matching. All of the limitations

Page 140


mentioned above are related to an RFID tag antenna, apart from the limitation of chip

sensitivity.

The following investigation in this section details challenges in designing UHF RFID

tag antenna and is based on two co-authored papers: “Analysis of constraints in small

UHF RFID tag design” [97] and “Small UHF RFID label antenna design and limita-

tions” [98].

• Size Limitation

An electrically small antenna is often defined as an antenna with length less than1

10 of wavelength at its operating frequency.

For HF (13.56 MHz), 110 of wavelength is approximately 2.2 m while for UHF (860

- 960 MHz), 110 of wavelength is approximately 31 - 35 mm. Since the size of an

RFID tag always depends on the size of its antenna, and for most applications it

is not feasible to have a large RFID tag, an RFID tag antenna is often small in size

and considered as an electrically small antenna.

A small antenna often has small radiation resistance. From (6.26), a small r will

result in a high quality factor, Q. A high Q means small bandwidth.

Ref. [99] shows that many electrically small antennas can be matched to a 50 Ω

line. However, the quality factor of the antenna is said to be not significantly

changed.

• Bandwidth Limitation

According to Bode and Fano, a fundamental limitation on impedance matching

takes the form [100]:

∫ ∞

0ln

1|Γ|dω ≤ π

RC(6.32)

where Γ is the reflection coefficient of the load and its assumed lossless matching

network with respect to the source impedance RS, and R and C is the resistance

and capacitance, respectively, that comes from the parallel RC load.

Page 141


To have a wide matching frequency band, a maximum mismatch outside ∆ω (the

band of interest) is required.

The best utilisation of πRC is to keep |Γ| constant (say at |Γ|inband) over the band

∆ω, and unity outside this band as shown in Fig. 6.7.

Figure 6.7. Reflection coefficient for best utilisation of λRC .

Based on the case shown in Fig. 6.7, (6.32) becomes

|Γ|inband ≥ e−1

2∆ f RC (6.33)

It is found in practice that with the chip input resistances and capacitances nor-

mally encountered, good power transfer can be obtained over the bandwidth of

interest for RFID. More details can be found in [97].

• Gain Limitation

An RFID tag antenna is categorised as small antenna. An efficient small RFID tag

antenna is normally assumed to have a gain of 1.5. For an operating RFID reader

antenna the maximum transmitted power that is allowed under FCC regulations

is 1 W into an antenna with a maximum gain of 6 dBi.

• Polarisation Loss

Page 142


From Section 6.2, the polarisation efficiency between a circularly polarised RFID

reader antenna and a linearly polarised RFID tag antenna is 0.5 (or -3 dB). How-

ever, it is almost impossible to have a perfect circularly polarised antenna. An

antenna is more likely to be elliptically polarised and depending on the axial ra-

tio, the maximum polarisation loss will increase with respect to the -3 dB of a

circularly polarised antenna. Ref. [101] shows that with a 3 dB axial ratio (2:1

ratio), the maximum polarisation loss will increase to 4.77 dB.

• Impedance Matching

An RFID chip is always capacitive in nature due to its reservoir capacitor in its

rectifying circuit. A matching antenna to a capacitive RFID chip must be induc-

tive in nature.

In the case of HF, the inductance of the HF antenna is tuned to resonate with the

capacitance of the HF chip at the frequency of operation. The inductance needed

can be provided by a multi-turn loop antenna. More discussion can be found in

Section 6.10.

In the case of UHF, Eq. (6.22) is used to compute the power transfer from the

antenna to the chip. The best case is when the impedance of the UHF antenna is

the conjugate match of impedance of the UHF chip.

However, a simple dipole that is usually less than half-wavelength in length nor-

mally has a capacitive input impedance [85]. A matching network is required

to match the dipole to the RFID chip. Commercial UHF RFID tag antennas nor-

mally have a matching network very similar to a conventional T-match network,

which can be found in [85].

The discussions of impedance matching will be included in the presentation of

designed tag antennas throughout this chapter. An examination on the effect of

impedance mismatch on the maximum power transfer between a tag antenna

and a tag chip can be found in Section 9.6.

Page 143

6.6 RFID Chips

6.6 RFID Chips

This section describes the common RFID chips used in experiments to test the perfor-

mance of an RFID tag antenna in terms of readability and its maximum read range.

In the early stage, HF ICs were obtained from Texas Instruments (RF-HDT-SJME-G1

Tag-it HF-I transponder IC) [102] and NXP-Phillips (I Code SLI SL2ICS20 Smart Label)

[103]. Both of these tag ICs conform to ISO 18000-Part 3 Mode 1 [33] and ISO 15693 [34].

The problem is that there is no way to attach an IC directly on to an antenna. Hence

a discussion was held with a technical group from Defence Science and Technology

Organisation (DSTO), Australia. It was discovered that a solder bump can be used to

connect electrically an IC to a thin metal film as shown in Fig. 6.8.

HF Chip

Solder1.5 mm Bond

Metallic Pad Ceramic

3 mm

0.38 mm

Few micron

Solder

Figure 6.8. The proposed chip bonding on to a thin metal strip.

However, due to the cost factor, the idea of bonding of ICs on to a conductive strips

was abandoned. As an alternative, ICs from existing tags were reused. The tag of

choice is Texas Instruments Tag-It HF-I C04, as shown in Fig. 6.9(a). An HF chip is cut

out carefully from a Tag-It HF-I C04 with a small portion of the metal film remaining

attached. The chip is then attached on to a prototyped antenna as shown in Fig. 6.9(b).

There is an electrical connection from the chip to the antenna through the metal film.

A special tape is used to stick the metal film and the antenna together. A thorough

discussion can be found in Section 6.8.

The chip reused from a Tag-It HF-I C04 will be known as the TI HF chip in this the-

sis. It has an input capacitance of 23.5 pF ±10% [102] at the operating frequency of

13.56 MHz.

Page 144


(a) A Tag-It HF-I C04

Tag Chip

(b) The chip of a Tag-It HF-I C04

reused on prototyped antennas

Figure 6.9. Reuse of HF RFID tag chip.

Another tag chip which can be reused is the ISD72128 or C220, originally designed

by Integrated Silicon Design Pty Ltd (an Australian company) and manufactured by

Chartered Semiconductor Manufacturing (Singapore) as shown in Fig. 6.10. However

this tag chip uses a proprietary (Tag Talk First) protocol and can only be interrogated

by proprietary RFID readers, also available originally from Integrated Silicon Design

Pty Ltd. The functionality of this TTF chip is as follows with respect to the induced

voltage at its input terminals: (1) At low voltage, the tag chip will generate a 100 µs

burst of 100 kHz square wave. This is an EAS (Electronic Article Surveillance) burst. (2)

At higher voltages the tag will generate a reply burst, 4 cycles of either 250 or 400 kHz,

depending upon the code written to memory. The sequence repeats at a duty cycle

of between 10 to 20 percent, depending upon excitation levels, 10 percent for high

excitations and 20 percent for low excitations.

Figure 6.10. A ISD72128 or C220 HF RFID tag.

There is a variant of HF tag chips with much higher input capacitance. For example the

NXP-Phillips SL1 ICS31 01 ICODE1 Label IC [103] has an input capacitance of 97 pF.

The advantage of having a higher input capacitance is that the antenna for the chip

Page 145

6.6 RFID Chips

can have fewer turns to reduce the overall dimensions, as the antenna does not need

to provide a high inductance to resonate with the input capacitance of the chip.

Also used was the ISD72128 packaged in a smart card module. The tag chip is as

shown in Fig. 6.11. The tag chip which is already mounted on the smart card module

is easy to solder on to an antenna.

Figure 6.11. 72128 TTF chip on smart card module.

Unlike the difficulties found in mounting the HF chip on to an antenna substrate, UHF

chips are readily available attached by thin metal straps. A special tape (Section 6.8) is

used to attached a UHF strap on to a UHF antenna.

An example of a UHF strap used in the fabrication of a UHF tag is as shown in Fig. 6.12,

which is the RI-UHF-STRAP-08 UHF Gen2 Strap from Texas Instruments [94]. It has a

input impedance of 380 Ω‖2.8 pF which is 9.7866 − j60.193 Ω at 920 MHz.

Figure 6.12. A Texas Instruments UHF strap.

Another type of UHF strap used in experiments is the C1G2 UHF strap from Alien

Technology (Fig. 6.13). It has an input impedance of 1.5k Ω‖1.3 pF which is 11.713 −j132.03 Ω at 920 MHz.

Page 146


Figure 6.13. An Alien C1G2 UHF strap.

6.7 RFID Readers

This section lists the common RFID readers used for the all the testing presented in

this thesis. HF readers operate at 13.56 MHz. UHF readers can be controlled to op-

erate within 860 to 960 MHz. Unless testing was carried in a controlled environment,

such as an anechoic chamber, the measurement results presented in this thesis for UHF

were carried out in adherence to Australian regulations, that is in the band of 920 to

926 MHz, at 4 W EIRP (Effective Isotropic Radiated Power).

6.7.1 HF Reader

• Fig. 6.14 shows an ID ISC.LR2000 HF reader and an ID ISC.ANT300 HF reader

antenna, both from FEIG Electronics. It supports ISO 15693, ISO 18000 Part 3

Mode 1.

• Fig. 6.15 shows an L120 HF reader and an HF AC201 reader antenna, both from

Gemplus Tag Australia, previously known as Integrated Silicon Design Pty Ltd.

It supports the proprietary TTF protocol. The suitable RFID tags are the ISD72128

(C220).

6.7.2 UHF Reader

• Fig. 6.16 shows an ID ISC.LRU2000 UHF RFID reader and an ID ISC.ANT.U250/250

UHF reader antenna, both from FEIG Electronics. It supports 18000 Part 6, EPC

C1G1 and EPC C1G2 tags.

Page 147

6.7 RFID Readers

1

23

(a) Reader

1

(b) Antenna

Figure 6.14. FEIG HF RFID Reader. (a) 1: Connection to reader antenna; 2: Optional connection

to a computer through RS232 interface; 3: Connection to a computer through a LAN

connection using TCP/IP protocol. (Connection to power supply not shown) (b) 1:

Connection to a reader.

123

(a) Reader

1

(b) Antenna

Figure 6.15. Gemplus HF RFID reader. (a) 1: Connection to reader antenna; 2: Connection to

a computer; 3: Connection to power supply. (b) 1: Connection to a reader. Note that

there is a ferrite core on the cable to reduce noise.

1

2

(a) Reader

1

(b) Antenna

Figure 6.16. FEIG UHF RFID reader. (a) 1: Connection to reader antenna; 2: Connection to a

computer. (b) 1: Connection to a reader.

Page 148


• Fig. 6.17 shows an ALR 9780 UHF RFID reader and an ALR 9610-BC reader an-

tenna, both from Alien Technology. It supports EPC C1G1 and EPC C1G2.

1

2

3

4

(a) Reader

3

(b) Antenna

Figure 6.17. Alien UHF RFID reader. (a) 1: Connection to a computer using serial connection; 2:

Connection to power supply; 3: Connection to reader antennas; 4: Unused connection

to antennas. (b) 3: Connection to a reader.

6.8 RFID Tag Fabrication

An RFID tag is the combination of an RFID tag antenna and an RFID tag chip, where

the tag chip is located across the input terminal of the RFID tag antenna.

For prototyping, the raw material for an RFID tag antenna is the Printed Circuit Board

(PCB). An example is the commonly used FR4 board. There are many types of PCB

boards uasable in the prototyping of the design RFID tag antenna. For the conductive

layer, copper and aluminium are the most common. For the dielectric layer, FR4 is the

most common.

There is a selection of choices for the thickness of the dielectric layer. The commonly

used ones are 1.6 mm and 0.2 mm. Also, the thickness of the metallic layer can be

chosen from within a few common values.

Most of the time, double sided PCB boards are needed for most of the antenna designs.

For the HF coil antenna, the bottom layer is needed for the underpass, which is linked

to the top layer through via holes.

Via holes are created using a rivet if a pad size of at least 1.3 mm in diameter is possible.

If, due to space constraints and no pad with at least 1.3 mm being available, a via hole

Page 149

6.9 Antenna Design and Simulation

is created by first drilling a hole using mini drill (Smallest possible diameter is 0.3 mm)

and a fine conductor is passed through the hole and soldered at both ends.

For the case of the UHF antenna, the bottom layer is always used for matching pur-

poses. The most common method is to have the bottom layer providing series or paral-

lel capacitance. No physical connection is needed between the top and bottom layers.

A typical HF coil antenna, with low fabrication cost, is a planar spiral strip antenna

on a dielectric substrate, as shown in Fig. 6.18. The dielectric substrate chosen in this

research work when fabricating an HF antenna is FR4, with relative permittivity of 4.4.

Figure 6.18. Typical HF antenna. Planar spiral in shape on a dielectric slab, normally FR4.

To attach a tag chip, 3MTM Z-Axis Electrically Conductive Tape 9703 is used. As sug-

gested by its name, this conductive tape only conducts in the Z-axis. It has a low

contact resistance . However it is not clear, from the specification sheet, what capac-

itance it may introduce into the circuit. A simple experiment was carried out to test

the effect of this conductive tape on an RFID tag. As shown in Fig. 6.19, a small part

of the HF coil antenna was removed. A small copper tape was then attached using

the Z-axis conductive tape to replace the removed HF coil track. A measurement us-

ing a Gemplus L120 reader shows that there is no degradation in performance. Hence

it is concluded that the Z-axis conductive tape has minimal impact on the RFID tag

antenna.


Page 150


Figure 6.19. Effect of using Z-axis conductive tape on an HF RFID tag.

6.9.1 Ansoft HFSS

High Frequency Structure Simulator (HFSS) is a 3D electromagnetic-field simulator

from Ansoft. This section documents the setup undertaken for all the simulations car-

ried out for RFID antenna design, for both the HF and UHF cases. This section is based

on the HFSS User Manual from Ansoft [104].

Ansoft HFSS uses the finite element method (FEM) to solve the simulation. In the

literature, there are other types of solver, which include the popular finite-difference

time-domain (FDTD) and the method of moments (MOM). This thesis does not inves-

tigate the difference between these solvers. A good description of the different solvers

or computational electromagnetic techniques can be found in [105].

Below are the common HFSS settings for all the simulations carried out for antennas

presented in this thesis:

Solution Type:

• Driven Modal Solution

For the calculation of modal-based S-parameters of passive, high frequency struc-

ture (including microstrip, transmission line). The scattering parameter solutions

will be expressed in terms of the incident and reflected powers of waveguide

modes.

Page 151


• Driven Terminal Solution

For the calculation of terminal-based S-parameters of multi-conductor transmis-

sion line ports. The scattering parameter solutions will be expressed in terms of

terminal voltages and currents.

• Eigenmode Solution

For the calculation of eigenmodes and resonances of a structure.

From experience, for the purpose of antenna simulation, the Driven Modal Solution is

the most appropriate choice. The Driven Terminal Solution is more suitable for circuit

analysis.

Excitation Point

Excitations in HFSS are used to specify the sources of electromagnetic fields and charges,

currents, or voltages on objects or surfaces in the design.

There are several types of excitation points in HFSS. The definition given by Ansoft for

the two common types are listed below:

1. Wave Port

Represents the surface through which a signal enters or exits the geometry.

2. Lumped Port

Represents an internal surface through which a signal enters or exits the geome-

try.

The difference between a wave port and lumped port is subtle. Both are represented by

a geometrical square or rectangle shape as an opening, which allows a signal to enter

or exit a structure (antenna in our case).

As suggested by the name, a wave port is often used to represent a waveguide or a

transmission line and is characterised by a characteristic impedance. In simulation a

wave port has to be defined at the edge of a structure.

Page 152


On the other hand, a lumped port can be characterised by a user defined complex

impedance and can be located internally. It is normally used for a microstrip structure.

Almost all of the antennas designed and presented in this thesis are fabricated on PCB

boards and have good resemblance to a microstrip structure. Hence, a lumped port is

chosen. The impedance of lumped port is set as 50 Ω instead of the impedance of the

RFID chip used.

A quick test showed that a change in the impedance of the lumped port will affect the

simulation results. For example, the input impedance of the antenna will change from

20 + j170 Ω at 950 MHz to 25 + j182 Ω at 950 MHz if the lumped port impedance is

changed from 15 Ω to 50 Ω. The reason a 50 Ω lumped port is chosen for simulation

is that the network analyser used for input impedance measurement is a 50 Ω system.

Hence, choosing 50 Ω will make the comparison between measurement results and

simulation results more meaningful and more accurate. More in-depth discussion on

measurement techniques and challenges is presented in Chapter 9.

Also, experience shows that simulation using a capacitive lumped port (for example

5 − j150 Ω to represent a capacitive UHF RFID tag) will result in much longer simula-

tion times.

The disadvantage of using a 50 Ω lumped port is that simulation results based on a

50 Ω lumped port may not be reflecting the actual situation when a UHF RFID tag is

connected to the antenna. To overcome this short-coming, most RFID tag antennas are

designed to have a structure with tuning flexibility. Since the difference in impedance

is not significant, a quick fine-tuning can optimise the performance of a RFID tag.

Boundary Condition

Similar to excitation types, there are several options for setting up a boundary for a

simulation. A boundary is required to specify the region of the problem. Without a

boundary, HFSS would not be able to confine the problem to a reasonable size.

The two common types of boundary for antenna simulation are as follow:

1. Radiation Boundary

Page 153


In HFSS, radiation boundaries are used to simulate open problems that allow

waves to radiate infinitely far into space, such as antenna designs. HFSS ab-

sorbs the wave at the radiation boundary, essentially ballooning the boundary

infinitely far away from the structure.

2. Perfectly Matched Layers (PML)

Perfectly matched layers (PML) are fictitious materials that fully absorb the elec-

tromagnetic fields impinging upon them. These materials are complex anisotropic.

Figure 6.20. PML layer in HFSS. The larger cubicle in the middle defines the working zone, and

is surrounded by PML layers.

In HFSS the radiation boundary must be at least quarter of wavelength away from the

radiating structure of interest. Also, a radiation boundary must be convex with respect

to a radiating source. This poses a lot of difficulties for the simulations of antennas

presented in this thesis. For example, at HF (13.56 MHz), a quarter of wavelength

is approximately 5.53 m. The antenna size is often in the size of up to 0.1 m. This

means in the simulation, the antenna must be located in a huge empty region, and the

simulation duration will increase significantly.

One may argue that an RFID HF antenna operates through inductive coupling and not

radiation. Hence a small radiation boundary is sufficient. However, HFSS computes

the power entering an antenna of interest and the power reflected from the antenna,

in order to compute the impedance of the antenna. With a radiation boundary smaller

than the recommended quarter wavelength, more power will be reflected, reducing

the accuracy of the simulation.

Page 154


One can compare a simulation to an actual measurement. Having a radiation boundary

in simulation is comparable to having absorbing material in an anechoic chamber for a

measurement. The absorbing material must be a certain distance away depending on

the frequency of operation of the antenna under test.

Also, the radiation boundary must be convex in shape with respect to the radiating

source. It is easy to fulfill this requirement if the antenna under test is a simple loop

antenna or a simple dipole. For a more complex antenna with an unsymmetrical or

irregular shape, it is very difficult to construct an appropriate radiation boundary.

6.9.2 Scripting in HFSS

Although the HFSS working environment allows 3-D modelling and HFSS supports

parametric optimisation, to construct spiral planar coil structure of many turns (> 4

turns) using HFSS is very cumbersome. The latest version of HFSS (version 10) sup-

ports modelling using a mathematical equation. However, an HF coil antenna is not

readily defined mathematically.

The alternative is to use the scripting function in HFSS to create the HF coil. The com-

prehensive HFSS guide is provided by Ansoft [106] and Virtual Basic Script (VBS) is

used as the programming language in the scripting of HFSS.

However, VBS lacks analytical tools. To write a VBS programme from scratch is time

consuming. Hence, it was decided to use MATLAB as the analytical tool. MATLAB

and HFSS are linked together to produce the desirable results. MATLAB will compute

the geometry of the structure using mathematical equations and compile HFSS script

using Virtual Basic Script (VBS). The link between MATLAB and HFSS is shown in

Fig. 6.21.

The foundation of the work is largely based on the API produced by V. C. Ramasami

[107]. Extensive extensions and modifications of the API have been made to suit the

work of the author.

The complete MATLAB script can be found in Appendix D. This MATLAB script de-

fines the function “makecoil”, which takes in the following parameters:

Page 155


MATLABScript

VB ScriptHFSS

Simulation

generate run

export result

Figure 6.21. Linking MATLAB and HFSS. MATLAB script is written to generate VB script. The

VB Script is then fed into the HFSS simulator to construct the antenna model and

set up simulation parameters. The simulated results are then exported from HFSS to

MATLAB for analysis and plotting.

• turn: This is to specify the number of turns of the coil antenna.

• L: The length of the inner-most coil.

• L1: The width of the inner-most coil.

• thick: The thickness of the copper film on top (or other conductive material). It is

normally set as 0.018 m.

• height: The thickness of the dielectric board. Normally a FR4 board of 1.6 mm

thickness is used.

• w: The width of the HF coil track.

• g: The gap between the coil tracks. A gap of 0 mm will result in a rectangular

patch as all the coil tracks will be merged together.

• save loc: The location and the filename of simulated results. The files will contain

both the S11 and the input impedance of the simulated antenna with respect to a

default frequency range of 860 to 960 MHz.

An example of an automatic generated HF coil model is as shown in Fig. 6.22.

A comparison between Table 6.2 (results from HFSS simulation) and Table 6.3 (results

from Section 6.10.2 based on self and mutual inductance concepts) and shows that

when the number of turns is small (< 6 turns), the simulated results are very close

to the computed results. However, as the number of turns increases, the difference

between these two results increases.

Page 156


(a) 2-loop Coil (b) 10-loop Coil

Figure 6.22. The auto generated loop using VBS.

Table 6.2. Inductance values for spiral planar coil with different numbers of turns, with L =

20; L1=40; w=3; g=1. The simulation of reactance considers both the inductance and

capacitance of a coil antenna. An effective inductance is obtained though Z = 2π f L.

Turns Reactance at 13.56 MHz (Ω) L (µH)

1 13.7120 0.16106

2 50.0168 0.58748

3 108.090 1.26950

4 191.700 2.25170

5 315.140 3.70160

6 504.470 5.92540

7 772.480 9.07340

8 1203.50 14.1360

9 2155.40 25.3170

10 4351.00 51.1060

In simulation, the impedance of the antenna is computed, not the inductance value. As

an approximation, the impedance is transformed back to the inductance value, using

Z = 2π f L.

6.9.3 Plotting in HFSS

HFSS can provides results in several forms. The most common results obtained are the

input impedances of the simulated results.

Page 157


In the simulation of an HF RFID antenna, the resonant frequency of an antenna can be

observed from the plot of real and imaginary parts of the simulated impedances. At

resonance the real part of the impedance will increase towards infinity and the imagi-

nary part will be zero.

In the simulation of the UHF RFID antenna, the input impedance can be presented

in the plot of impedance versus frequency on a Smith Chart. Also available from the

simulation of the UHF RFID antenna is the bandwidth of the antenna. To obtain the

bandwidth of the antenna itself is easy, by using a 50 Ω lumped port for simulation,

and plotting the |S11| of the simulation results. However, this |S11|, as shown in Trace

1 in Fig. 6.23 shows the bandwidth of the combination of the tag antenna and the

50 Ω source, but does not offer any insight into the bandwidth of the RFID tag alone.

Moreover, the bandwidth of an RFID tag is affected by both the tag antenna and the

chip attached to the antenna.

0.85 0.90 0.95 1.00 1.05-25.00

-20.00

-15.00

-10.00

-5.00

0.00

Freq [GHz]

1

2

Figure 6.23. Plots |S11| curve in HFSS. (Y-axis in dB). Trace 1: |S11| from the simulation which

represents the |S11| which would be seen if the antenna is connected to a measuring

instrument using a 50 Ω cable; Trace 2: Eq. (6.22) in dB terms, representing the

suggested way to plot a |S11| curve in HFSS, with an assumed chip impedance.

Normalising the simulation results using the target impedance of the designed antenna

is one way suggested by HFSS to obtain information on tag bandwidth. HFSS has a

built in post processing function to normalise results based on a new port impedance

Page 158


value. The complete arithmetic for the normalising process can be found in [104]. In

simple term, HFSS computes a new |S11|, or |Snew| using:

|Snew| =

∣∣∣∣∣Z0

1+S111−S11

− Znew

Z01+S111−S11

+ Znew

∣∣∣∣∣ (6.34)

where Znew is the target impedance of the designed antenna (or in other words, the

conjugate impedance of the chosen RFID chip.)

An example is shown in Fig. 6.24. Trace 1 is showing the plot of 20 log10 |Snew|, which

is obtained after the post processing normalisation.

0.85 0.90 0.95 1.00 1.05-40.00

-30.00

-20.00

-10.00

0.00

Freq [GHz]

1

2

Figure 6.24. Plots in HFSS showing the normalised |S11| curve. (Y-axis in dB). Trace 1:

Normalised |S11|; Trace 2: Eq. (6.22) in dB terms.

However, a problem lies in the fact that normalisation only works sensibly when the

new target impedance is purely real, which is not the case here as the target impedance

(a chip impedance), is both resistive and capacitive. As can be seen from Fig. 6.24, the

arithmetic of normalisation results in very low |Snew| value and it is not possible to

determine the bandwidth of the RFID tag. Also, this method assumes a constant RFID

chip input impedance, which in fact is frequency dependent.

The solution is to use (6.22) as the foundation in determining the bandwidth of an RFID

tag. Firstly, from Section 6.6, an equivalent circuit for a UHF RFID tag chip consists of a

Page 159

6.10 HF RFID Antenna Design

resistor in parallel with a capacitance (in the example shown in this section, the resistor

is assumed to be 1 kΩ and the capacitor is assumed to be 1 pF). Using this equivalent

circuit, the resultant input impedance of a UHF RFID tag chip within a frequency band

can be computed. Note that the input impedance changes as the frequency increases.

By substituting this computed input impedance of the chip and also the simulated

antenna impedance from HFSS into (6.22), the ratio between the power loss and the

power available can be computed. This ratio plotted in dB and is shown as Trace 2 in

both Fig. 6.23 and Fig. 6.24.

The bandwidth of an RFID tag (tag antenna and tag chip) can then be obtained from

Trace 2. Generally, the bandwidth is defined as the frequency range where the |S11|curve is below the -3 dB line. Within this bandwidth, an RFID tag is expected to func-

tion properly according to its specification.

6.9.4 ISO-Pro

ISO-Pro is a software program from T-Tech [108] to control the T-Tech’s quick proto-

typing system.

T-Tech’s quick prototyping system is actually a milling machine. It mills away un-

wanted metal from a PCB, and the remaining metal forms the designed antenna.

The 3-D antenna model is drawn in HFSS for simulation. The same simulation model is

used to redraw the designed antenna in Protel (Protel is a PCB layout software, which

can export a designed layout to ISO-Pro with ease).

From HFSS, the antenna model is exported in the format of AutoCAD DXF format.

However, a DXF format is a 2-D format. Hence, if a model has several layers, several

DXF files are needed.


This section presents the design of an HF RFID antenna. An HF RFID system operates

at 13.56 MHz. An HF RFID antenna is designed so that together with an RFID tag

Page 160


chip, it resonates at the operating frequency. Firstly, this section discusses the use of

a simple loop antenna as an HF RFID antenna. In the later part, various novel HF

antenna designs for various applications are shown.

The literature, [95, 109, 110, 111] provide guidelines in designing HF RFID antennas.

The fundamental idea in designing an HF antenna is always to have the designed an-

tenna providing sufficient inductance to tune with the HF chip capacitance and hence

to have a resonant point at the operating frequency. This section is not meant to over-

lap with existing literature. This section touches lightly on the theory and focuses

more on the application-driven engineering challenges: from understanding a unique

RFID application, initial antenna simulation, to antenna prototyping and finally an-

tenna testing.

6.10.1 A Simple Loop Antenna

A simple loop antenna acts as a magnetic dipole. The formulae used and the detailed

design procedures are described by Balanis [87] and Bernhard [92]. Optimisation of an

HF antenna is discussed by Chen [112] while Lee offers a design note on RFID antenna

circuit design [109].

An HF electrically small circular loop antenna with radius a and wire radius b [87] has

a radiation resistance given by:

Rr = 20π2(βa)4 Ω (6.35)

and inductance,

L = µoa[

ln(

8ab− 2

)](6.36)

For a square or rectangular (not circular) loop antenna, its characteristics can be esti-

mated using the area encompassed by the loop:

πr2 = a × b (6.37)

Page 161


where r is the radius of an equivalent circular loop antenna, a and b are the width and

the length of the rectangular or square (where a = b) respectively.

6.10.2 A HF Planar Spiral Coil Antenna

The characteristics of a simple one-turn planar coil antenna can be estimated using the

characteristics of a simple loop antenna. The equivalent of a simple one-turn planar

coil antenna as a simple loop antenna can be found by using the table of conductor

geometrical shapes and their equivalent circular cylinder radii [85]. A width of the

planar coil, a, is related to the radius of the cylindrical coil, ae, though the relation

ae = 0.25a.

For a multi-turn coil antenna where the coil maintains a fixed radius, the total induc-

tance can be estimated using:

Ltotal = n2 × L (6.38)

where Ltotal is the total inductance, L is the inductance of one loop, and n is the number

of loops.

However, it is impossible to have a multi-turn planar coil antenna with fixed radius.

Hence to predict the total inductance of a multi-turn planar coil antenna, a more com-

plex method from [113] is applied. (6.39) to (6.45) and their explanations are adapted

from [113].

This complex method is based on the idea of inductance per unit length. However, in

a physical world, the idea of inductance per unit length does not make much sense as

this idea implies that a short straight wire segment has a certain amount of inductance.

Hence, any statement about the inductance per unit length or the inductance of a wire

segment must always be recognised as being only a contribution to the true inductance

(self or mutual) of a complete circuit(s) (in our case is a loop antenna consists of more

than 4 wire segments) and that the complete inductance will consist of the sum of the

so-called inductances of all wire segments and so-called mutual inductances between

all segments. A comprehensive discussion on this matter can be found in Appendix B.

Page 162


Bearing that in mind, the exact expression for the inductance of a straight conductor

(or wire segment) is given as:

L = 0.002l[

ln(

2lGMD

)− 1.25 +

AMDl

+(µ

4

)T]

(6.39)

where

• L = inductance (µH)

• l = conductor length (cm)

• GMD = Geometric Mean Distance of the conductor cross section. GMD is defined

as the distance between two infinite thin imaginary filaments whose mutual in-

ductance is equal to the mutual inductance between the two original conductors.

For a single conductor, GMD is defined as the distance between two imaginary

filaments normal to the cross section, whose mutual inductance is equal to the

self-inductance of the conductor. The computation of GMD is often lengthy but

for common cross sectional shapes, the GMD can be obtained from the literature.

In the case of planar coil, where the cross sectional shape is a very thin rectangle,

the value for GMD is given by 0.22313× (a + b), where a and b are the width and

height of the cross sectional rectangle.

• AMD = Arithmetic Mean Distance of the conductor cross section. AMD for a

single conductor is the average of all possible distances within the cross section.

In the case of planar coil, where the cross sectional shape is a very thin rectangle,

the values for AMD is given as 13 .

• µ = conductor relative permeability.

• T = frequency-correction parameter. Although T can be considered 1 for mi-

crowave frequencies, we still assumed a value of 1 in HF. The argument for that

is out of the study of this research and is an assumed value.

To reduce (6.39) for our inductance calculation, we use T = 1, µ = 1, and consider the

copper strips (of the antenna loop) as having a very thin rectangular cross sectional

shape (i.e. b → 0):

Page 163


L = 0.002l

ln[

2la + b

]+ 0.50049 +

[(a + b)

3l

](6.40)

The aim is to compute the inductance of a planar spiral antenna. The total inductance

is contributed by both self inductance and mutual inductance by all the segments of

the coil. The self inductance can be obtained by calculating the self inductance of each

segment using (6.40) and summing all the self inductances of the segments.

The computation of mutual inductance is more complex, as there are positive and neg-

ative mutual inductances. Positive mutual inductance is caused by two parallel con-

ducting segments with the same current direction while negative mutual inductance

is caused by two parallel conduction segments with different current directions. We

assume that there is negligible mutual inductance between two perpendicular con-

ducting segments.

Hence the total inductance of a planar coil antenna is given by (6.41):

LTotal = Lsel f + M+ − M− (6.41)

where Lsel f is the total self inductance, M+ is the total positive mutual inductance of

all segments (in nH), M− is the total negative mutual inductance of all segments (in

nH).

Between a pair of parallel conducting segment, there exists either a positive or negative

mutual inductance. The magnitude for both positive or mutual inductance can be

calculated using (6.42):

M = 2lQ (6.42)

where l is the length of the segment (in cm), and Q (do not confuse with quality factor,

Q) is the mutual inductance parameter calculated using (6.43).

Q = ln

⎧⎨⎩

(l

GMD

)+

[1 +

(l

GMD

)2] 1

2⎫⎬⎭ −

[1 +

(GMD

l

)2] 1

2

+(

GMDl

)(6.43)

Page 164


However, if the two parallel conducting segments are not the same in length (which is

the case of a planar spiral antenna), the mutual inductance between them has to be

calculated using (6.44) and this equation is best explained by Fig. 6.25.

2MA,B = (Mb+p + Mb+q) - (Mp + Mq) (6.44)

where MA,B represents the mutual inductance between segment A and segment B, and

Mlength (such as Mb+p or Mp) can be computed using (6.42) and (6.43).



Figure 6.25. Computation of the mutual inductance between two planar wire

segments.

In the figure, planar wire segment A is j in length while planar wire segment B is m in

length. The difference length between planar wire segment A and B is (p+q). (Based

on [113].)

In the computation of the inductance of the planar spiral antenna, it is assumed that

the length of p and q are the same and equation (6.44) is then reduced to:

MA,B = Mb+p - Mp (6.45)

This planar spiral antenna must provide sufficient inductance so that the reactance

provided by the inductance can cancel the reactance provided by the chip capacitance

at the frequency of interest (which is 13.56 MHz for HF operation). If we assume a

chip capacitance of approximately 1 pF in series, we would need a series inductance

of 140 µH in series, calculated using:

(6.46)

Page 165


This amount of inductance is difficult to achieve using a spiral planar coil structure.

The normal practice is to increase the capacitance instead, by attaching a parallel ca-

pacitor to the chip’s input capacitance. However, at this stage, it is not known what

the impact of the additional parallel capacitance to the self-resonance frequency of the

spiral planar coil will be.

The computation of positive and negative mutual inductances is complex and lengthy.

Hence a MATLAB script is written to compute these mutual inductances. The complete

code is attached as Appendix E.

Together with the self inductance (refer to (6.40)), the total inductance can be com-

puted. Table 6.3 shows some of the results for spiral planar coil on different dimen-

sions:

Table 6.3. Inductance values for various spiral planar coils. The length of the inner-most coil

= 20 mm; the width of the inner-most coil = 40 mm; the width of the track = 3 mm;

the gap between the coil tracks = 1 mm. Lsel f is the self inductance of spiral coils; M−

is the total negative mutual inductance; M+ is the total positive inductance; LTotal is

the total inductance of spiral coils.

Turns Lsel f M− M+ LTotal (µH)

1 0.2093 0.0333 0.0352 0.2112

2 0.4631 0.1413 0.3177 0.6395

3 0.7513 0.3391 0.8781 1.2903

4 1.0747 0.6426 1.7247 2.1568

5 1.4339 1.0683 2.8746 3.2402

6 1.8294 1.6330 4.3485 4.5448

7 2.2617 2.3543 6.1691 6.0765

8 2.7313 3.2499 8.3604 7.8419

9 3.2387 4.3381 10.947 9.8476

10 3.7842 5.6377 13.954 12.101

Page 166


6.10.3 An HF Antenna for a New Wine Closures

This section describes the design process for an RFID HF tag antenna for a special wine

closure. This wine closure is a new invention from Zork [114]. It is made of plastic

with a thin metal foil on top to optimise its functionality as a wine closure, which is

to preserve the quality of the wine. Fig. 6.26 shows a model of the wine closure of

interest. An RFID tag is confined in the space shown in the figure.

Plastic Cover

Thin Metal Sheet

Space for TagPlastic Stopper

Figure 6.26. A Zork wine closure model. This diagram is not to scale and does not represent

an actual Zork wine closure. This diagram is to indicate the location of the space

available for RFID tag.

Figure 6.27. RFID tag for wine closure.

The aims of embedding an RFID tag into the wine closure from the point of wine re-

tailer include product authentication, tracking, and anti-theft and anti-counterfeiting.

From the view of customers, they can scan a wine bottle embedded with an RFID tag

in some specific area to obtain extra information about the wine they have just chosen,

Page 167


such as the the origin of the wine, ingredients, awards, and even recommendations

from wine critics around the world.

The aim is to have a multi-turn coil antenna to provide inductance to resonate with

the HF chip input capacitance. Besides the space limitations, it is expected that the

thin metal foil will reduce the performance of the RFID tag. The final design of the HF

tag antenna is shown in Fig. 6.27. Table 6.4 shows the details and read range perfor-

mance of the designed HF RFID tag for Zork wine closure. It should be noted that the

maximum read Range is limited by the reader antenna size [95].

Table 6.4. Performance of HF RFID tag for Zork.

Material Copper wire

Dimension (mm) 25.0 (h) 7.5 (r)

Chip ISD72128 or C220

Reader ISD L120

Read Range (mm) 150 (Best)

6.10.4 An HF Antenna for a Pigs

This section describes the design process for an RFID HF tag antenna for a pig ear tag.

The pig ear tag project is sponsored by the Pork CRC, Australia. The design of a UHF

RFID tag antenna for pig ear tagging is presented in Section 6.11.4. The pig ear tag

is a relatively large project, when compared with the Zork wine closure tag antenna

design.

The common frequency used in livestock tagging is the Low Frequency (LF) band ,

spanning 125 - 135 kHz. LF RFID operation is standardised by ISO 11784/85 [51, 52].

The LF RFID system normally cannot handle a dense tag environment. Also, LF re-

quires large antenna components and hence is difficult to implement and is susceptible

to electrical noise, which HF can handle [20].

Page 168


Although a new enhanced version of an LF chip is available [115] with anti-collision

capability, the traditional LF tag which complies with ISO 11784/85 does not offer such

anti-collision capability. Without anti-collision, a single LF tag must only be present in

the reader antenna interrogation zone at any one time. For practical real life applica-

tions, animals are forced to pass through (or enter the interrogation zone) one by one.

Where this is practicable is when identifying larger animals, such as cattle or sheep.

Examples include cattle and sheep farming.

However, in pig farming, a problem arises. Small baby piglets are smaller in size as

compared with cattle. It is not easy to herd piglets through a gate mounted with an

antenna. Hence it is the intention of this research work to investigate the feasibility of

using either an HF or UHF RFID tag (Section 6.11.4) for the tagging of pigs, by using

the development of pig ear tag as the case study.

Livestock ear tags for cattle deployed in Australia must be accredited by National Live-

stock Identification System (NLIS), Australia. Examples of these livestock ear tags with

embedded RFID tag can be found in [116, 117], which is ISO 11784/85 compatible.

These tags can be recovered at slaughterhouse and reused, and can withstand haz-

ardous environments. A cattle livestock ear tag before any RFID tag is embedded

is shown in Fig. 6.28(a). One part of the ear tag has a hole in the middle to allow

both parts to be pinned together on the ear of cattle livestock. The space to attach

an RFID tag is hence restricted to a circular disc with a hole in the middle as shown

in Fig. 6.28(b). The outer diameter of the space available is 28.3 mm, while the inner

diameter is 12.0 mm.

In the early design stage, antennas were designed to be fitted into the casing shown in

Fig. 6.28. In the later stage of the research, the new industrial partner has provided a

new casing for encapsulation, as shown in Fig. 6.29. The available space of this new

casing has an outer diameter of 27.8 mm and inner diameter of 14.5 mm.

The initial design incorporated the following key features of an HF tag:

• A multi-turn loop to provide the required inductance.

Page 169


Part 1

Part 2

(a) (b)

Figure 6.28. Casing for livestock RFID tag. (a) Two parts of a cattle livestock tag so that it

can be pinned on the ears of cattle livestock. (b) The possible space for RFID tag

allocation on a cattle livestock tag: outer diameter of 28.3 mm and inner diameter of

12.0 mm.

Figure 6.29. Tag encapsulation casing from Leader Products. The dimension of spacing avail-

able: outer diameter of 27.7 mm and inner diameter of 14.5 mm.

• Allow as much space as possible in the middle of the coil to have a high induc-

tance value.

• An underpass and two via holes to complete the loop.

The antenna has 6 loops on one side of an FR4 substrate (1.6 mm thick) as shown in

Fig. 6.30. The formation of the multi-turn loop is achieved by connecting several almost

circular ring tracks together. In the simulation software, the circular shapes are formed

by polygons, which are visibly obvious in Fig. 6.30. Also, the connecting part is bent

and is not smooth.

Page 170


Figure 6.30. First version of a pig tag with bent tracks to connect all the circular tracks

together..

An alternative way to model the multi-turn loop antenna is to use a spiral structure. A

spiral multi-turn loop is shown in Fig. 6.31. There are no bend as compared to Fig. 6.30.

However, the simulation of a spiral model takes a comparatively longer time. Also,

similar to the circular shape, it is noticeable that the spiral track is formed by polygons

and the edges are not smooth.

Figure 6.31. Second version of a pig tag using a spiral track.

Hence, a spiral is not chosen as the final choice for the modelling of an HF pig ear tag.

Circular shapes are also avoided to prevent the polygonal representation. The solution

Page 171


is rather peculiar. Rather than using a polygonal circular shape, a higher order polygon

is defined. The choice of using a high order polygon can be seen in Fig. 6.32, which is

also the figure for the first prototype HF pig ear tag.

Figure 6.32. First prototype of a pig tag.

After the simulation, the antenna was sent for fabrication. A serious problem was

encountered during the fabrication. As shown in Fig. 6.33, both the gap between the

coil track and the width of the track itself of the HF antenna are very small. This often

results in coil track short-circuiting. Also, the uneven track increases the loss resistance

of the antenna. This will reduce the quality factor and reduce the performance of the

tag.

12

Figure 6.33. Problem of uneven track thickness when prototyping antenna with thin tracks.

(1) Thinner tracks, (2) Thicker tracks.

Page 172


After a satisfactory antenna was fabricated despite the problem mentioned above, a

simple test was carried out to check the resonance point of the fabricated antenna. A

small capacitor was soldered on to the feed point of the antenna to represent the input

capacitance of a HF RFID chip, as shown in Fig. 6.34.

Figure 6.34. First version prototype of a pig tag. No HF chip is attached. A capacitor of 100 pF

is used instead to represent the input reactance of a HF chip.

The added capacitor had a capacitance value of 100 pF, equal value to the HF chip input

capacitance of 100 pF (Section 6.6), to mimic an HF chip. The tag, HF antenna with the

added capacitor was then tested for its resonant frequency. It was then discovered that

the resonant frequency was 14.71 MHz.

This shows that with slight modification (slight increment in inductance) of the HF

tag antenna design shown in Fig. 6.34, the HF tag antenna will operate in the desired

frequency of 13.56 MHz. A suitable HF chip is the NXP-Phillips SL1 ICS31 01 ICODE1

Label IC [103] which has an input capacitance of 97 pF.

It should be noted that the test of tag resonant frequency by substituting an HF chip

with a capacitor is purely an early indication of whether the design approach is suitable

or not; this is achieved by measuring the resonance frequency. A test using an actual

HF chip is needed to confirm the functionality of the tag. As mentioned before in

Section 6.6, there are no available HF chip straps which resemble the UHF chip straps.

As discussed in Section 6.6, all the HF chips used in experiment are recycled chips from

existing HF tags. Hence, it is not feasible to use recycled HF chips at the beginning of

the design cycle. After simulation and fabrication, an easy way to test the design is to

Page 173


use a tag chip with Tag Talk First (TTF) compatibility (ISD72128, refer to Section 6.6).

A tag (same design with the one shown in Fig. 6.34 attached with a tunable capacitor

and a TTF HF chip is shown in Fig. 6.35.

Figure 6.35. Pig tag with TTF compatibility.

1

2

3

Figure 6.36. Testing a pig tag with TTF compatibility. (1) To a signal generator, (2) To a

reader antenna, (3) To a spectrum analyser.

Page 174


No RFID reader is required to test a TTF RFID tag. Once a TTF RFID tag is powered

by a CW wave at the correct frequency, it will start generating a signal, depending on

the induced voltage at its input terminal (See Section 6.6 for more details). Hence the

testing environment is as shown in Fig. 6.36, where a signal generator, a spectrum anal-

yser, an antenna and a three port directional coupler. The connection to the directional

coupler is as follows:

• Out: Connected to Signal Generator.

• In: Connector to the antenna of HF reader.

• Cpl: Connected to spectrum analyser to monitor the response from a tag.

This test shows a read range of approximately 0.2 m. The results prove that the tag is

working as an RFID tag. No fine-tuning was carried out to optimise the performance

of the tag as it was then decided to pick an HF chip with lower input capacitance. This

is due to the fact that the input capacitance value has great impact on the cost of an

HF chip as the chip will need bigger area to provide the necessary capacitance. Hence

it is most desirable to have a tag antenna design which can function on a lower input

capacitance HF chip, such as the 23.5 pF input capacitance RF-HDT-SJME-G1 Tag-it

HF-I transponder IC from Texas Instrument [102].

An HF antenna needs to provide sufficient inductance through its coil turns to resonate

with the tag chip capacitance at the frequency of operation, which is 13.56 MHz. In the

case of a size constraint and reduced HF chip input capacitance, an HF tag would not

have sufficient inductance, and extra capacitance would be added by making the HF

antenna double sided and introducing capacitance through the two overlapping plates.

To avoid the difficulties in fabrication (uneven tracks as shown in Fig. 6.33 caused

by the closely spaced tracks), the tracks are designed to be further apart as shown in

Fig. 6.37. Simulation was carried out using Ansoft HFSS to predict the performance of

the designed tag antenna.

The simulation results are as shown in Fig. 6.38. The resonant point of the HF an-

tenna alone is at approximately 40.2 MHz. The measurement results on a fabricated

prototype antenna show a resonant point at the frequency of 41.3 MHz.

Page 175


Figure 6.37. Pig tag with external capacitance added.

0.00 10.00 20.00 30.00 40.00 50.00-20000.00

-10000.00

0.00

10000.00

20000.00

30000.00

Freq [MHz]

Figure 6.38. Simulation results for the pig tag in Fig. 6.37 at HF with a thickness of 1.6 mm.

(Red curve is the resistance curve and blue curve is the reactance curve, Y-axis in Ω).

To have a high capacitance value, it is desirable for the tag to be as thin as possible to

reduce the distance between the overlapping coils and plates so that the capacitance

between the parallel plate in the middle and also the capacitance between the upper

and the lower rack of the HF coil are increased. The thickness of the FR4 was reduced

from 1.6 mm to 0.8 mm. The simulation results are as shown in Fig. 6.39. The simulated

resonant point is at approximately 29.2 MHz. The measurement results on a fabricated

prototype antenna show a resonant point at the frequency of 28.6 MHz.

Page 176


0.00 10.00 20.00 30.00 40.00 50.00-10000.00

0.00

10000.00

20000.00

Freq [MHz]

Figure 6.39. Simulation results for the pig tag in Fig. 6.37 at HF with a thickness of 0.8 mm.


Several simulations, antenna fabrications and tests were carried out to optimise the

performance of the tag. The final design of the HF antenna is shown in Fig. 6.41. The

thickness of the substrate has been reduced to 0.2 mm. Also, the coil tracks separa-

tion cannot be increased due to the size limitation. The chip that was used was the

Texas Instrument Tag-it HF-I Standard Transponder IC [102], which is based on the

ISO/IEC 15693 and ISO/IEC 18000-3 standards, and has an nominal input capacitance

of 23.5 pF.

12

3

Figure 6.40. Final version prototype of a pig tag. (1) Multi-turn coil to provide inductance, (2)

TI HF chip, (3) Tunable capacitance plate for fine tuning.

Page 177


The front and the back of the un-tuned final design is shown in Fig. 6.41.

(a) Front view (b) Back view

Figure 6.41. HF RFID Tag to be embedded in a livestock ear tag.

For testing purposes, the used ID ISC.LR2000 HF reader and ID ISC.ANT300 reader

antenna, both from FEIG Electronics were used. The width and length of the reader

antenna are both approximately 0.33 m. In HF operation, the maximum possible read

range is proportional to the size of the reader antenna [95]. As a rule of thumb in real

life deployment, the maximum read range is approximately equal to the dimension of

the reader antenna. A read range of 0.34 m is obtained in free space and of approx-

imately 0.32 m when attached to a human hand, which simulates the environment

where a tag is attached to the pig’s ear. To increase the read range, we can increase the

size of the reader antenna in a way presented in [95].

After encapsulation, a simple test was performed to certify that an encapsulated tag

was in fact made water proof. This simple test did not focus on any potential degrada-

tion of tag due to the presence of water. A tag was randomly selected and immersed in

water as shown in Fig. 6.43. The tag is constantly monitored over a long period as the

water container was positioned in the middle of an operating HF reader antenna (ID

ISC.ANT300/300-A) controlled by an HF reader (ID ISC.LR2000), both from FEIG Elec-

tronic. The immersed tag was functioning as usual which confirm the encapsulation

process was successful in preventing water damage to an RFID tag.

The deployment of HF RFID tags was carried out together with UHF RFID tags, where

the design of UHF RFID tags for pig tagging can be found in Section 6.11.4. The com-

plete report of RFID deployment in piggery can be found in Appendix G. Fig. 6.44

Page 178


(a) Before encapsulation (b) After encapsulation

Figure 6.42. RFID tags before and after encapsulation process. (a) HF (top) and UHF (bot-

tom) RFID tags embedded in the casing shown in Fig. 6.29. Discussion of UHF pig

tag can be found in Section 6.11.4. (b) An encapsulated RFID tag viewed from (from

left) front, side and back.

Figure 6.43. An encapsulated tag immersed in water to test for readability and water proof

capability. Tag remains functioning even after all the water has evaporated.

shows the overall RFID system setup in the pig pen in the piggery. Fig. 6.45 shows

the arrangement of HF and UHF RFID readers in protective casings in the piggery.

Fig. 6.46 shows the orientation of HF and UHF tags on pig ears, and Fig. 6.47 shows

the snapshot when pigs are feeding in an RFID enabled feeder.

Part of this section has been published in [118] and [119].

6.11 UHF RFID Antenna Design

This section presents the design of a UHF RFID antenna. A UHF RFID tag operates

within 860 MHz to 960 MHz around the world. A UHF RFID tag antenna is designed

Page 179


Figure 6.44. Overall RFID system setup in piggery. Feeder on the left with UHF antennas and

feeder on the right with HF antennas.

Figure 6.45. HF and UHF RFID readers in piggery. Readers setup in piggery: HF (bottom) and

UHF (top) RFID readers in protective boxes.

Figure 6.46. Orientation of HF and UHF tags on pigs ears. A pig was tagged with one HF tag

(left in picture) and one UHF tag (right in picture).

to have a conjugate match with a UHF RFID chip to allow maximum power trans-

fer. Firstly, this section discusses the common commercial UHF RFID tags and their

common features. In the later part, various novel UHF antenna designs for various

applications are shown.

Page 180


Figure 6.47. Picture on pigs feeding in an RFID pig feeder. Pigs feeding at both the UHF (left

in picture) and HF (right in picture) feeders.

UHF tags, as discussed before, are very susceptible to the environment. Hence, the

design of UHF RFID tag antenna is very diverse, largely depending on its applica-

tion. There are numerous research papers on the design of UHF antennas for various

applications [120].

The design of UHF RFID antenna is not of the main research interest of the author. As

it can be seen later, most of the designs of UHF tag antenna come from joint research.

However, it is an important learning process as the design of a dual-frequency antenna

in Chapter 7 will require the design of UHF RFID tag antenna.

6.11.1 A Common UHF RFID Tag

A common commercial UHF RFID tag has very distinctive features. It is either a dipole,

or a small variation of the dipole, and has a simple matching network, most commonly

a matching network very similar to a T-Match is used to transform the capacitive na-

ture of a dipole into inductive to conjugately match to the input impedance of a UHF

RFID tag chip.

Also, common UHF RFID tags are very thin and flexible. They can readily be applied to

a pallet , or carton as used in retail supply chains. They can have a read range of more

than 10 m, in some regulatory regions, using the air interface specified by EPCglobal

C1G2 [35].

Page 181


6.11.2 Simple Planar Dipole

This section presents the fundamental principles of operation of a simple planar dipole.

A dipole or electric dipole is the dual of a loop antenna or magnetic dipole. As sug-

gested by the name, an electric dipole antenna is sensitive to electric field while a mag-

netic dipole antenna (loop antenna) is sensitive to magnetic field. As a common name,

an electric dipole is often known as a dipole while a magnetic dipole is known as a loop

antenna. A dipole and a loop antenna have very similar radiation patterns (doughnut

shape) when correctly aligned.

A half wave length dipole is theoretically purely resistive (zero reactance). However in

actual implementation, a half wave dipole has to be shortened by about 4-10 percent to

present a purely resistive impedance to the feed point. From [85], the input impedance

of a half wave length dipole, Zin, is equal to 73 + j42.5 Ω with a directivity of 1.643.

In theory, as mentioned above, a half wave dipole is purely resistive. A dipole shorter

than a half wave dipole is always capacitive. Increasing the size of dipole past the

purely resistive point will push the impedance into the inductive region. An inductive

dipole without any matching network is generally significantly longer than a half wave

length.

The UHF band for RFID operation around the world is from 860 MHz to 960 MHz. A

half wave length dipole for this frequency band is between 0.3125 m and 0.3488 m. A

typical UHF RFID tag is normally around 0.1 m in length, and hence, a dipole without

any matching network is not feasible to be used as a UHF RFID tag antenna.

Some of the popular matching networks for a dipole are known as the Omega and

Gamma matching networks. However, an Omega matching network can only match

loads less than 50 Ω resistive by increasing the resistance. An Omega matching cannot

step or transform antenna feed resistance downwards when using capacitors. Also,

both Omega and Gamma matching networks cannot match capacitive antenna loads.

A variation of the dipole is the folded dipole. The directional characteristics of a folded

dipole are the same as those of a simple dipole. However, the reactance of a folded

dipole varies much more slowly as the frequency is varied from resonance. Because of

Page 182


this the folded dipole can be used over a much wider frequency range than is possible

with a simple dipole. For RFID applications, the dipole, folded or not, is always planar.

6.11.3 A UHF Antenna for Sheep

This section presents a UHF antenna for sheep tagging. This section is an extract from

a co-authored paper “A small passive UHF RFID tag for livestock identification” [121].

Similar to pig tagging as discussed in Section 6.10.4, animals such as cattle and sheep

are tagged for purposes such as disease control, breeding management and also stock

management. Also similar to the pig tagging, where a passive HF RFID tag is attached

on a pig’s ear, sheep tagging uses a small passive UHF RFID tag to be attached on a

sheep’s ear.

The current RFID technology for livestock tagging, including both pig tagging and

sheep tagging, operate in the low frequency (LF) region [116]. However, as discussed

in Section 6.10.4, LF tags can only be read at close range and may not perform well

when multiple tags are simultaneously present in the interrogation field.

It is believed that the HF RFID tag designed for pig tagging as presented in Sec-

tion 6.10.4 can be used in sheep tagging as well. However, the focus of the research

on sheep tagging is on UHF operation. As compared to HF tags, UHF tags not only

give better read range, but also support higher data rates.

The design of the UHF RFID sheep tag is based on a loop antenna, contrary to the

common dipole built UHF RFID tag. The main reason is that a loop will provide in-

ductance useful in the matching to the capacitive input impedance of a tag chip. Also,

a loop antenna does not have sharp edges and can be fitted on the sheep’s ear relatively

easily.

This sheep tag is designed to operate in the frequency of 915 MHz. (Note: This fre-

quency is not normally allowed in Australia. The test was carried out at that frequency

but in a controlled environment.). The UHF chip used is the Alien C1G1 while the

reader used is the Alien UHF RFID reader.

Page 183


Figure 6.48. Sheep ear tag. (a) Front view; (b) Rear view of the sheep ear tag

The front and rear view of the final product is shown in Fig. 6.48. The matching net-

work is provided by the copper strips on the rear side, providing series and shunt

capacitance to the circuit. These copper strips can be trimmed to fine tune the circuit,

during development, so that the antenna will resonate at the desired frequency, which

is 915 MHz. This antenna is designed to operate in the USA, but with a simple tuning

(trimming of copper strips on the rear side of the antenna), it can operate in Australia

conforming to the local regulations.

The target impedance of the designed antenna is 6.5371 + j133.44 Ω, as compared to

the chip impedance of 17.36 + j149.2 Ω at 915 MHz. The maximum measured read

range was about 1.42 m.

However, as predicted by the radiation pattern, when the tag was aligned with its

axis parallel with that of the interrogator antenna, the reading distance was only a few

centimetres. A simple example of the possible the deployment of this tag for sheep

Page 184


tagging is to locate the reader antenna on top of a door arch where the tagged animals

pass through, with the direction of the reader antenna facing down.

6.11.4 A UHF Antenna for Pigs

This section is based on a published paper “Investigation on the deployment of HF and

UHF RFID tag in livestock identification” [118]. The background discussion of RFID

tags for pigs is presented in Section 6.10.4.

Two different UHF tag designs have been considered in our study. Both UHF tags have

been designed for operation at 923 MHz in Australia. The tag chips used for both tags

are from Alien Technology [122], and have impedance of approximately 20 − j141 Ω at

923 MHz.

25

mm

2 mm

copperstrip

substrate

chip

FRONT

13

mm

1 mmseriescapacitor

shuntcapacitor

BACK

x

y

(a) Tag 1(b) Tag 1: radiation

pattern

28m

m

2 mm

13

mm

0.5 mm

chip

x

y

(c) Tag 2(d) Tag 2: radiation

pattern

Figure 6.49. Two UHF pig tags.

The first UHF tag design consists of a circular loop antenna with a two element match-

ing network to match the impedances of the tag antenna and tag chip. More detailed

design steps of this tag can be found in [121]. However, the tag has been made slightly

larger in diameter compared to that presented in [121] to suit the application studied in

Page 185


Table 6.5. Read range for UHF tags.

UHF Tag 1 UHF Tag 2

Tag in free space (m) 1.00 0.48

Tag attached to hand (m) 0.20 0.27

this research. The structure and dimensions of this tag is shown in Fig. 6.49(a). The tag

is made using a thin FR4 board with substrate thickness h = 0.36 mm and relative di-

electric permittivity εr = 4.4. This tag design utilises both sides of the FR4 board. From

Fig. 6.49(a), the front view is the circular loop antenna and the back view is the imple-

mentation of the matching network that consists of a series and a shunt capacitor. The

tag antenna structure is simulated using Ansoft HFSS and the simulated impedance is

10 + j148 Ω, which is the conjugate of the tag chip impedance. The radiation pattern

of the antenna is as shown in Fig. 6.49(b), with maximum directivity of 1.65 dB.

The second UHF tag design considered consists of a curved electric dipole antenna

with an inductance track across the dipole for impedance matching purposes. The

structure and dimensions of this tag is shown in Fig 6.49(c). This tag is made using an

FR4 board with substrate thickness h = 1.6 mm and relative dielectric permittivity εr =

4.4. This tag design is single-sided. The tag antenna structure is simulated using Ansoft

HFSS and the simulated impedance is 1 + j152 Ω. As can be seen, the impedance of

the tag antenna structure is not exactly the conjugate of the tag chip impedance and

hence, the tag antenna and chip impedances are not perfectly matched. However,

the inductance track across the dipole does provide sufficient inductance to tune with

the capacitance of the tag chip. The radiation pattern of the antenna is as shown in

Fig. 6.49(d), with maximum directivity of 1.83 dB. As can be seen, although the general

shape of the radiation pattern of the first and second UHF tags are almost the same,

the maximum directivity of these two tags occur at different directions with respect to

the plane of the tags. Both UHF tags are tested and the read range results are shown in

Table 6.5.

The complete report of RFID deployment in piggery can be found as Appendix G.

Page 186


6.11.5 A UHF Antenna for Beer Kegs

This section presents a UHF antenna for a Beer Keg. This section is an extract from a co-

authored paper “A Small Passive UHF RFID Tag for Metallic Item Identification” [123].

The tag presented here is not only for the deployment of RFID in beer keg tagging but

the main purpose is to illustrate how an RFID tag can be designed to be placed on a

metallic surface. The idea of this tag design is to fully exploit the fact that magnetic

field is doubled at a metallic surface. The loop antenna is oriented in such a way as

to couple to the magnetic field at the metallic surface. The final design is as shown in

Fig. 6.50.

Tag antenna

Tag

chip

Width (W )rec Length ( )Lrec

Height

(H )rec

x y

z

Figure 6.50. Structure of the RFID tag with a rectangular loop antenna.

The simulated antenna impedance is shown in Table 6.6. The read range of 0.83 m

is obtained when a reader is operating at reduced power. The reason for testing at re-

duced operating power, is that this application is likely to be used with hand held RFID

readers, and to conserve battery power, the RF output power is limited to 500 mW. It

is expected the tag can be read up to 1.3 m when a reader is operating at maximum

allowable power.

Table 6.6. Read range for UHF tag for beer keg.

Antenna Impedance Peak Directivity

Free Space 0.26 + j91 Ω 1.5 dB

Above Metal 0.53 + j91 Ω 6.5 dB

Page 187

6.12 Conclusion

6.11.6 A UHF Antenna for New Wine Closures

This section describes the design of a UHF antenna for a wine closure. The wine closure

of interest is the same as presented in Section 6.10.3 and illustrated in Fig 6.26.

Figure 6.51. RFID tag for wine closure.

The functionality of the UHF tag for the wine closure is also the same with the the

HF tag for wine closure as discussed in Section 6.10.3. The final design is as shown in

Fig. 6.51 and the details and result of actual testing is presented in Table 6.7.

Table 6.7. Performance of UHF RFID tag for the Zork wine closure.

Material Copper strip

Dimension (mm) 6.0 (h) 7.0 (r)

Chip Alien Class 1 Generation 1

Reader Alien ALR-9780

Read Range (mm) ≈ 300

6.12 Conclusion

This chapter presents the fundamentals of HF and UHF tag designing, novel HF and

UHF RFID tag antenna designs for new applications, antenna modelling, simulation

Page 188


and prototyping, and tag fabrication process. This provides a strong foundation to

advance into the design of dual-frequency antenna for RFID application, which is es-

sentially a single antenna operating as both a HF antenna and a UHF antenna.

Page 189

Chapter 7

Dual-Frequency RFIDAntenna

WITH good read range and fast data rates, UHF Radio

Frequency Identification (RFID) is being deployed in the

supply chain to uniquely identify each item. However,

UHF transmission power is easily absorbed by ionised liquids, such as wa-

ter. Hence, normally, in an environment with lots of liquid products, an

HF RFID system is chosen instead. This chapter investigates the feasibility

of designing a dual-frequency (HF and UHF) RFID tag antenna, with fre-

quency ratio of up to 70, to embrace the benefits offered by both the UHF

and HF RFID systems. Furthermore, this chapter shows that with care-

ful design, a single feed dual-frequency RFID tag antenna can be achieved.

A miniaturised version of a dual-frequency RFID tag antenna is also de-

signed. Prototypes of the dual-frequency RFID tag antenna are presented

in this chapter, along with validating simulations and measurement results.

Page 191

7.1 Introduction

7.1 Introduction

As discussed in Chapter 2, Radio Frequency Identification (RFID) is a technique used

to identify objects by means of electromagnetic waves. An electronic code responding

label, also known as a tag, consists of an antenna and an integrated circuit. Upon

receiving any valid interrogating signal from any interrogating source, such as an RFID

reader, the tag will respond according to its designed protocol.

RFID has been recently applied into supply chain tracking systems, revolutionising

the conventional ways of how an object is tracked along the supply chain and how an

inventory system works. Not only does RFID bring benefits in saving costs, it also has

the potential of increasing security levels along the supply chain, especially in genuine

goods authentication. To reduce the cost of RFID deployment, passive tags are used.

There are 4 common bands for RFID applications; these are the LF band (less than

135 kHz), the HF band (13.56 MHz), the UHF band (860 - 960 MHz) and the so-called

microwave (2.45 GHz). LF is not suitable for supply chain deployment, as to uniquely

identify every object in the supply chain, each tag has to bear a unique EPC number,

which must be at least 64 bits in length [4]. A low frequency operation, such as in

the LF band, would suffer very slow reading in a heavily populated tag environment.

Also, LF requires large antenna components and hence is difficult to implement and

is susceptible to electrical noise, which HF can handle [20]. UHF and microwave tags

can offer comparatively very fast reading, but their performance will suffer more than

the other bands described above in the presence of liquid or metal [19]. It is very dif-

ficult to conclude whether HF or UHF is better for RFID application in supply chains.

This is because in any supply chain, there are many different scenarios, and in some

HF is better and in the others, UHF outperforms HF operation. Table 7.1 summarises

comparison between HF and UHF as presented in [15, 19, 24].

To embrace both the advantages contributed by HF and UHF, this chapter investigates

the feasibility of designing a dual-frequency antenna for an RFID tag. The next sec-

tion reviews current technology on dual-frequency antennas, and their shortcomings

in designing an RFID HF and UHF dual-frequency antenna. Design aims ensuring

Page 192

Chapter 7 Dual-Frequency RFID Antenna

Table 7.1. HF versus UHF in RFID operation.

HF UHF

Read Range ≈ 1 m Up to 10 m

Cost (in large volume) Medium Low

Read Rate Low High

Metallic Area Bad Bad

Liquid Surrounding Better Worse

that the antenna is feasible for real life application are presented in Section 7.3. Sec-

tion 7.4 describes the design process in detail, including the problems encountered

and the solution of those problems using simulation results. Section 7.5 presents the

actual measurement results for the designed dual-frequency antenna, together with

some discussions and comments.

Section 7.6 begins the redesign of the novel dual-frequency antenna as presented in

Section 7.4, with the emphasis on miniaturisation. Also included in this section, are

the characteristics of the novel dual-frequency antenna (Section 7.4) retained for the

new miniaturised antenna design. The new or improved characteristics are discussed

in Section 7.7. Section 7.8 explains the test procedures used to fine-tune and verify

the functionality of the new miniaturised dual-frequency antenna while Section 7.9

shows the final design with its dimensions. This chapter is based on two published

paper [77, 124].

7.2 Current Dual-Frequency Antenna Design

The intended dual-frequency RFID antenna is planned to function at 13.56 MHz (HF)

and also within the UHF band for RFID operation (860 - 960 MHz). If using the high-

est frequency in the UHF band, which is 960 MHz, the frequency ratio for this dual-

frequency antenna is slightly more than 70.

Page 193

7.3 Design Aims

Wong has undertaken extensive research on dual-frequency microstrip antennas, with

the frequency ratio ranging from 1.8 - 4.9 [125]. However, the frequency ratio of this

technique is too low to be applied to the intended RFID dual-frequency antenna. A

common aperture, dual-feed dual-frequency antenna for 900 MHz and 60 GHz was

presented by Menzel [126]. The frequency ratio achieved was approximately 70. How-

ever, the antenna had dual feed points.

One of the antennas which has a high frequency ratio and used in RFID is the one

produced by IPICO [127]. Under the IPICO dual-frequency system, the RFID reader

transmits a signal using the LF spectrum (125 kHz to 135 kHz) in order to power the

tag and the tag uses the HF spectrum (3 to 30 MHz) to transmit its signal back to the

reader. The tag produced by IPICO is using LF and HF, while the aim of this chapter is

to combine both HF and UHF.

7.3 Design Aims

The aim is to design, simulate and fabricate a dual-frequency RFID antenna, which can

demonstrate the following characteristics:

1. Antenna impedance is equal to the complement of the input impedance of the RFID chip

at UHF operation.

There is no available dual-frequency RFID chip to be used as the guidelines for

the design of dual-frequency RFID tag antenna. However, in a hypothetical

way, it is predicted that a dual-frequency RFID tag chip will have a similar in-

put impedance at UHF as the input impedance of a commercially available UHF

RFID tag chip (such as those presented in Section 6.6).

With a dual-frequency antenna having the impedance complement to a chosen

RFID chip at UHF, maximum power can be transferred from the antenna to the

RFID chip. The RFID chip used in the fabrication process has an approximate in-

put impedance of 17− 150j Ω at RFID UHF band. Hence the final antenna design

must have an input impedance of approximately 17 + 150j Ω. In actual fact, the

chip impedance will change within the RFID UHF band, but in this chapter the

Page 194


UHF antenna is designed to be matched at the highest RFID operating frequency

(960 MHz). It is believed that if an antenna can work at the highest frequency

(having the highest frequency ratio), the antenna can be tuned in a reasonably

easy way to operate at the lowest frequency (860 MHz).

2. A resonance point at HF.

Similar to point 1, where is no available dual-frequency RFID chip to be used as

the guidelines for the design of dual-frequency RFID tag antenna at UHF, there is

no guideline for the design of HF functionality. Again, in a hypothetical way, it is

predicted that a dual-frequency RFID tag chip will have a very small (negligible

level) of input capacitance at HF. This is not the case for a conventional HF chip.

From the HF chip examples from Section 6.6, the HF chip can have an input

capacitance from 23.5 pF up to 97 pF.

Hence, the designed dual-frequency antenna must have sufficient inductance

and capacitance to have a resonant point at 13.56 MHz. Also, in order to have

a high quality factor, the resistive losses must be kept as low as possible.

3. A single feed antenna.

The final antenna design is planned to be attached to commercially available

RFID chips. Hence a single feed antenna is desired to avoid any modification

required to the current RFID chips’ structure, where only a single feed point ex-

ists. Although there is no existing dual-frequency RFID tag chip available, it is

envisaged that a single feed dual-frequency RFID tag chip will have a lower cost

than a multi-feed dual-frequency RFID tag chip.

4. Reasonable antenna size and cost.

The material used for fabrication is FR4, with relative permittivity, εr, equal to 4.4.

The antenna area should not exceed 14400 (mm)2. However, it is not the focus at

the initial stage to minimise the size of the antenna. Hence, the first novel design

may not be of the smallest possible size. The design of a miniaturised version of

a dual-frequency RFID tag antenna can be found in Section 7.6 onwards.

Page 195

7.4 Dual-Frequency Antenna Design


Through brainstorming, three main methodologies to design a dual-frequency antenna

were identified:

1. Merging an HF antenna and a UHF antenna.

2. Making an HF antenna operate in UHF.

3. Making a UHF antenna operate in HF.

This chapter will focus on the first methodology, while Chapter 8 will present the re-

search results from the other methodologies.

The design philosophy used to design a dual-frequency antenna (HF and UHF), is to

design an HF and a UHF antenna separately as in Section 7.4.1, and merge these two

antennas together with a single feed. However, since these two antennas are joined

together in parallel, the UHF antenna will affect the HF operation of the HF antenna,

and vice versa. A few experiments were carried out to verify the merging idea and are

presented in Section 7.4.2 and Section 7.4.3.

7.4.1 Independent HF and UHF Antenna Design

For operation at HF, a multi-turn planar loop antenna is designed. The calculation of

the inductance of a multi-turn planar loop antenna is done using [113], which includes

the computation of the positive and negative mutual inductances. From f0 = 12×π

√LC

,

with an intended resonant frequency of 13.56 MHz, we know that a large inductance

in required. The problem is at the later stage, when this loop antenna is merged with

the UHF dipole, the resonant frequency will change. Hence, the number of turns of

the multi-turn loop is not fixed at this stage, and will only be finalised during the fine-

tuning stage after the merging of the HF and the UHF antennas.

An example of a generic multi-turn planar loop antenna is shown in Fig. 7.1. For oper-

ation at UHF, a dipole is used. An example of a generic half wave-length (of 960 MHz)

Page 196


dipole is shown in Fig. 7.2. Note that the feed points of the antennas are represented by

small rectangular boxes in those figures. This is the case for all the antennas presented

in this thesis.

Figure 7.1. A generic HF coil antenna. The multi-turn loop provides the required inductance.

Figure 7.2. A generic UHF dipole antenna.

Page 197


7.4.2 Quick Feasibility Test

To prove the idea of merging an HF antenna together with a UHF antenna, an HF an-

tenna is merged with a UHF antenna in HFSS to observe the feasibility and to form the

foundation of building a dual-frequency RFID tag antenna which will operate accord-

ing to specifications.

Generic models from Section 7.4.1 are used. A simple four-loop planar coil antenna is

used as the HF antenna and a simple dipole is used as the UHF antenna. Two via holes

and a underpass are required for the connectivity of the simple four-loop planar coil

antenna. The model of this dual-frequency antenna is shown in Fig. 7.3.

The width of the track is arbitrarily chosen as 3 mm with the gaps between the HF

planar coil loops chosen as 1 mm.

The antennas are interconnected by a transmission line as shown in Fig. 7.3. The length

of the transmission line is 80 mm, with the HF antenna connected to one side, and the

UHF antenna on the other side. This length includes the 3 mm track of both the HF

and the UHF antennas at the junctions where these two antennas are connected to the

transmission line. The length was chosen to be approximately λ4 , which is 78 mm (for

a frequency of 960 MHz). Although the permittivity of the substrate will affect this

length, it is not considered at this stage, but will be discussed in detail in later stages.

The length of dipole including the gap is 132 mm. The choice of length is based on

the fact that a practical half-wave dipole is shorter than its theoretical half-wavelength

which is 156 mm (at a frequency of 960 MHz). Again, the effect of the substrate is not

taken into account at this stage.

This initial simulation of a dual-frequency antenna with the proposed method of merg-

ing two different antennas together showed promising results. Fig. 7.4 shows a reso-

nant point at HF (32.5 MHz). Although the obtained resonant point is not exactly at

the intended frequency of 13.56 MHz, the addition of inductance or/and capacitance

will provide the correct resonant frequency. Fig. 7.5 shows that the resistance is higher

than the target value while the reactance is capacitive (the target reactance should be

inductive). Using a matching network, the impedance characteristics at UHF can be

Page 198


Figure 7.3. First version of a dual-frequency antenna. A generic HF coil antenna as shown in

Fig. 7.1 and a generic dipole as shown in Fig. 7.2 are connected together with a single

feed point.

0.00 10.00 20.00 30.00 40.00 50.00-10000.00

-5000.00

0.00

5000.00

10000.00

15000.00

Freq [MHz]

Figure 7.4. Impedance of the first version of dual-frequency antenna at HF. (Red curve is

the resistance curve and blue curve is the reactance curve, Y-axis in Ω). The fre-

quency response at HF of the antenna shown in Fig. 7.3. The resonant point occurs at

32.5 MHz.

easily matched to the tag chip impedance for the intended UHF frequency band as

shown in Section 7.4.4.

This quick feasibility test shows that the merging of two independently designed an-

tennas using a transmission line and a single feed is possible. Fine-tuning is required

Page 199


860.00 880.00 900.00 920.00 940.00 960.00-100.00

-50.00

0.00

50.00

100.00

150.00

Freq [MHz]

Figure 7.5. Impedance of the first version of dual-frequency antenna at UHF. (Red curve is

the resistance curve and blue curve is the reactance curve, Y-axis in Ω). The frequency

response at UHF of the antenna shown in Fig. 7.3. The antenna is capacitive across

the UHF band of interest.

for this dual-frequency antenna to operate in the desired frequency bands. A simple

test was carried out (Section 7.4.3) to check the tunability of this antenna.

7.4.3 Tunability Test

The first version of the dual-frequency antenna (Fig. 7.3) was then modified slightly to

test its tunability and is named second version of the dual-frequency antenna (Fig. 7.6)

for easy referencing. This quick tunability test is to demonstrate whether the character-

istics of the dual-frequency antenna can be changed in a predicable way when simple

modifications are applied.

In this test, the dipole is bent to reduce of the overall size of the dual-frequency an-

tenna. Also, the length of the transmission line linking the HF and UHF antennas is

varied (10, 20, 30, 40, and 80 mm), to observe the changes in impedance. An example of

the simulation model used is shown in Fig. 7.6 where the dipole is bent and the length

of the transmission line is 10 mm.

Page 200


Figure 7.6. Second version of dual-frequency antenna. The overall structure of the first version

of dual-frequency antenna is redesigned, with the shortening of the transmission linking

the HF and the UHF antenna, and the bending of the UHF dipole to reduce space.

An example of the graph showing the results obtained through simulation are shown

in Fig. 7.7, which corresponds to the simulation of the model in Fig. 7.6. This impedance

curves looks similar to the impedance curves of the first version dual-frequency an-

tenna (Fig. 7.5). The slight difference is the the reactance curve which has been shifted

up, and in parts of the frequency band, it has become inductive where it was previously

capacitive.

• Prediction of Impedance

Since the HF and the UHF antennas are connected in parallel, it might be possible

to predict the overall impedance if the individual impedance of both the HF and

UHF antennas are known. This test is to examine the coupling effect between the

HF and UHF antennas when placed near to each other.

The idea is to assume the antenna impedances are lumped and a simple parallel

circuit can be formed and the combined impedance, Zcal, can then be calculated

using Zcal = Zdipole||Zcoil, where Zdipole and Zcoil are the simulated impedances

of the individual UHF antenna and HF antenna respectively. If there is negli-

gible coupling effect between these two antennas, Zcal should be equivalent to

the simulated impedance for the dual-frequency antenna, Zsimu. However, the

Page 201


0.85 0.90 0.95 1.00 1.05-400.00

-200.00

0.00

200.00

400.00

600.00

Freq [GHz]

Figure 7.7. Impedance of the second version of dual-frequency antenna at UHF. (Red curve is

the resistance curve and blue curve is the reactance curve, Y-axis in Ω). The frequency

response at UHF of the antenna shown in Fig. 7.6. The reactance curve shows that

it is possible to push the reactance curve into the inductive zone with minimal design

modification.

requirement of a transmission line to link the HF and the UHF antennas together

will affect the results. This will be taken into consideration when making a con-

clusion on this simple test.

Table 7.2 shows the calculated combined impedance of a dual-frequency antenna

when given individual impedances of an HF antenna and a UHF antenna while

Table 7.3 compares the calculated combined impedance of a dual-frequency an-

tenna with the simulated impedance of the same dual-frequency antenna. Fig. 7.8

shows clearly the accuracy of the calculated impedance value as compared to the

simulated impedance based on the values shown in Table 7.3.

It can be seen from Fig. 7.8 that the calculated impedances are close to the sim-

ulated impedances, following a similar pattern. The inaccuracy occurs mostly

around the resonant zone (between 0.85 - 0.95 GHz), and can be improved if

more frequency points are considered.

Page 202


Table 7.2. Calculation of combined impedance based on individual impedances. Zdipole and

Zcoil are the simulated impedances of the individual dipole and coil antenna used to form

the dual-frequency antenna shown in Fig. 7.6. Zcal is obtained by summing in parallel

Zdipole and Zcoil.

Freq. Re(Zdipole) Im(Zdipole) Re(Zcoil) Im(Zcoil) Re(Ztotal) Im(Ztotal)

(GHz) (Ω) (Ω) (Ω) (Ω) (Ω) (Ω)

0.850 162.8325 122.4763 44.5804 -79.2678 71.2344 50.7455

0.875 205.5780 141.7485 106.0059 -19.0146 80.2015 -4.0880

0.900 260.7286 153.8072 114.7911 -130.8109 130.1086 51.7748

0.925 329.0677 151.2801 55.9386 -97.2257 75.9360 71.7810

0.950 406.5400 123.7083 44.1677 -60.3592 49.4806 49.2760

0.975 479.8166 61.5165 47.0725 -24.9616 44.3721 20.3141

1.000 526.5993 -33.9802 70.6423 -6.4963 62.2915 5.5255

1.025 528.5103 -142.8569 67.1498 10.2945 60.5988 -6.5156

1.050 487.4053 -236.7669 83.2981 50.2712 80.6469 -34.7300

Nonetheless, this experiment confirms that the total impedance of a single feed

dual-frequency antenna is dependent on both the impedance of the combined an-

tennas following simple parallel circuit theory. The small difference between Zcal

and Zsimu is caused by the addition of the transmission line and slight coupling

between the HF and the UHF antennas.

• Test of Transmission Line Length

As mentioned before, a transmission line is used to connect the HF and the UHF

antenna together with a single feed point to form a dual-frequency antenna. This

second quick feasibility test examines how the length of the transmission line can

affect the input impedance of the dual-frequency antenna. This transmission line

of interest is shown in Fig. 7.9.

The single feed point is to be located at one of the original feed points as shown

in Fig. 7.9(a). The decision to choose either point A or point B to be the feed point

is based on the following arguments:

Page 203


Table 7.3. Comparison between calculated impedances and simulated impedances. Zcal is

the computed impedance of the dual-frequency antenna shown in Fig. 7.6, by summing

in parallel the impedance of the individual dipole and coil antenna used to form the dual-

frequency antenna shown in Fig. 7.6. Zsimu is simulated impedance of the dual-frequency

antenna shown in Fig. 7.6.

Freq. Re(Zcal) Im(Zcal) Re(Zsimu) Im(Zsimu)

(GHz) (Ω) (Ω) (Ω) (Ω)

0.850 71.2344 -50.7455 181.7573 39.9317

0.875 80.2015 4.0880 178.9783 121.3906

0.900 130.1086 -51.7748 345.1479 -215.3287

0.925 75.9360 -71.7810 93.2314 -137.3622

0.950 49.4806 -49.2760 60.5289 -67.6305

0.975 44.3721 -20.3141 56.5010 -22.3317

1.000 62.2915 -5.5255 74.0738 10.1240

1.025 60.5988 6.5156 65.7329 17.8289

1.050 80.6469 34.7300 80.8109 75.9376

– Consider the case where point A is chosen as the single feed point. If the

UHF antenna acts as an open circuit at HF, the transmission line needs to be

as short as possible, so that the open circuited UHF antenna would not in-

terfere with the HF antenna at HF. If the UHF antenna acts as a short circuit

at HF, the length of the transmission line needs to be λ4 (HF) (approximately

5.5 m), which is physically too large to be implemented and attached on

most of the item. At UHF, the HF antenna will interfere with the UHF an-

tenna, and depending on the length of the transmission line, it could be very

difficult to fine-tune the antenna to obtain the matching impedance.

– Consider the case where point B is chosen as the single feed point. If the HF

antenna appears to be an open circuit at UHF, the transmission line must

be as short as possible so that it will still looks as an open circuit even with

the addition of the transmission line. If the HF antenna appears to be an

Page 204


-300

-200

-100

0

100

200

300

400

0.85 0.9 0.95 1 1.05

Freq (GHz)

Imp

eda

nce

(O

hm

)

Re(Simu)

Re(Cal)

Im(Simu)

Im(Cal)

Figure 7.8. Graph showing the calculated impedance and simulated impedance of the second

version of dual-frequency antenna. The calculated impedance, using the summation

of parallel impedances, has similar pattern with the simulated impedance. The difference

in results is caused by the coupling effect between the HF coil antenna and the UHF

dipole.

short circuit at UHF, the transmission line needs to be λ4 (UHF) (approxi-

mately 78 mm at 960 MHz and approximately 87 mm at 860 MHz), which

is realisable as an RFID tag antenna. At HF, the UHF antenna can be used

to provide additional capacitance to the HF antenna in obtaining a resonant

point at 13.56 MHz.

l

Coil Antenna

Dipole Antenna

A B

(a)

Bl

Merging of Antenna(Feed Point at B)

(b)

Figure 7.9. Location of feed point. (a) A transmission line is added between the two antennas.

(b) Feed point located at B.

Page 205


From these arguments, it is more sensible to design a dual-frequency antenna

with the feed point located at B (Fig. 7.9(b)). Hence feed point B is chosen, with

the length of the transmission line, l, equals λ4 (UHF). The loop antenna (for HF)

is to be designed to have low impedance at UHF, to be transformed into high

impedance using the λ4 (UHF) transmission line, so it will not interfere with the

UHF dipole. The UHF dipole should be capacitive at HF and, as mentioned be-

fore, will contribute the capacitance to the HF antenna.

As discussed when determining the feed point location, the length of the trans-

mission line is always fixed to λ4 , in order to transform a low impedance to high

impedance, and vice versa. However, even though it is believed at this stage that

the HF antenna will have a low impedance at UHF, it might not be as low as de-

sired, so that when transformed it will not become very high. In this simple test

of transmission line length, it is hoped that a small variation in the length will

result in a small and predictable change in impedance so that in any case where

fine-tuning is needed, the length of transmission line can be considered.

The simulated impedances of the dual-frequency antenna with different trans-

mission line length are plotted in a Smith Chart shown as Fig. 7.10. Simulations

were carried at various discrete frequency points and hence the plots are jagged

and not smooth. However, Fig. 7.10 provides sufficient insight to shows that

there is a pattern in the changes of impedance when the length is varied.

Both the simple examinations on the tunability of the to be designed dual-frequency

antenna returned positive results. Hence, the focus will be diverted to the actual design

of this planned dual-frequency antenna, with single common feed point located at

point B (Fig. 7.9(b)).

7.4.4 Redesign of the UHF Dipole

In the previous section, two important facts have been shown: (1) A single-feed dual-

frequency antenna can be obtained by merging two independent antennas, and (2) an

adjustable transmission line can be used to control the overall impedance.

Page 206


10mm

20mm

30mm

40mm

80mm

Figure 7.10. Smith Chart showing the simulated impedance of the dual-frequency antenna

with different transmission line length. As can be seen, as the transmission line

length changes (from 10 mm to 80 mm, represented by different colours as shown

in the top left corner of the Smith Chart), the impedance will change accordingly,

following a vague but definite pattern.

However, as can be seen from Fig. 7.5 and Fig. 7.7, the dipole is not matched to the RFID

chip impedance. It is not easy, and even may be impossible, to vary just the dimension

of the dipole to obtain the needed input impedance to have a perfect match.

Hence, the UHF dipole is redesigned so that it operates in the frequency band as in-

tended in a matching condition. The frequency band for RFID around the world is

within 860 - 960 MHz.

The first design is shown in Fig. 7.11. It is just a normal dipole with a matching network

added. The matching network chosen is a modified version of the T-match as discussed

in [85].

The designed antenna is tuned carefully to have a good match with a UHF chip for

a large bandwidth. The design is shown in Fig. 7.12, and the impedance within the

Page 207


Figure 7.11. A UHF dipole with a matching network. The matching network is a modified

T-match network.

UHF RFID band is shown in Fig. 7.13. However, the real part is slightly higher than

expected. Also, the dimension of the dipole is too long in the x-axis direction.

Figure 7.12. A UHF dipole with tuned matching network. The matching network (A modified

T-match) can be fine-tuned to operate in the band of interest. In our case, it is tuned

to conjugately match to a UHF RFID tag chip at 960 MHz.

Hence, the two sides of the dipole are bent as shown in Fig. 7.14. When comparing with

Fig. 7.12, the length at the x-axis is reduced from 168 mm to 78 mm, at the expense of the

overall length in the y-axis. However, this will not be a problem, as the transmission

line has to be positioned along the y-axis as well. The impedance of this new design is

Page 208


0.85 0.90 0.95 1.00 1.05 0.00

50.00

100.00

150.00

200.00

250.00

Freq [GHz]

Figure 7.13. Impedance of the UHF dipole with tuned matching network. (Red curve is the

resistance curve and blue curve is the reactance curve). This is the impedance curves

of the UHF dipole shown in Fig. 7.12.

shown in Fig. 7.15. The resistance of this new design has been lowered to well below

20 Ω for the entire UHF band for RFID (860 - 960 MHz).

From Section 6.2, the power loss caused by mismatch can be obtained using:

PlostPavailable

=∣∣∣∣Zant − Z∗

cctZant + Zcct

∣∣∣∣2

(7.1)

where Zant is the input impedance of the antenna and Zcct is the input impedance of

the RFID chip used.

From Section 6.6, at 960 MHz, the input impedance of the C1G1 chip used is ap-

proximately 16 − j140 Ω. To compare with other C1G2 chips, the input impedance

is 10.764− j126.61 Ω for an Alien C1G2 chip, 9.0070− j57.806 Ω for a Texas Instrument

C1G2 chip. The design was carried out to match the C1G1 chip. At this stage the power

loss is computed to be around 55 % using (7.1), contributed by mismatch in both the

real and imaginary parts.

Page 209


Figure 7.14. A UHF dipole with bent sides. The dipole is bent to conserve space but with slight

reduction in performance.

0.85 0.90 0.95 1.00 1.05 0.00

50.00

100.00

150.00

200.00

Freq [GHz]

Figure 7.15. Impedance of the UHF dipole with bent sides. (Red curve is the resistance curve

and blue curve is the reactance curve, Y-axis in Ω). This is the impedance curves of

the antenna shown in Fig. 7.14.

7.4.5 Compatibility of Dipole in HF

The dipole shown in Fig. 7.14 has a reasonably good match with an RFID UHF chip

within the UHF band (Fig. 7.15). However, the matching network of the UHF dipole

has provided a short circuit path for HF. The initial idea to rectify this problem is to add

a series capacitance which acts like an open circuit at HF and acts like a short circuit

Page 210


at UHF. The additional series capacitance takes the form of a gap, with an underlying

track, as shown in Fig. 7.16.

Figure 7.16. A UHF dipole with bent sides and gap. The gap prevents the UHF matching

network from short-circuiting the HF antenna at HF.

To maintain a balanced current flow, it is better to reposition the gap so that the gap

is symmetric along the y-axis. The end result is as shown in Fig. 7.17, (top view in

Fig. 7.18). The patch at the bottom of the gap can be adjusted easily (in both length and

width) for fine-tuning purposes.

Figure 7.17. A UHF dipole with bent sides and repositioned gap. The gap is repositioned to

maintain balanced structure for a balanced current flow.

Page 211


Figure 7.18. Final version of the UHF dipole, with matching network, bent sides and gap.

The simulation results of the final design of the UHF dipole with matching network

(Fig. 7.18), shown in Fig. 7.19 and Fig. 7.20, show the impedance in the HF band

and UHF band respectively. At HF, the antenna has a relatively large impedance.

Though the real part is significantly smaller than the imaginary part, at 13.56 MHz,

its impedance is 57.4 - 4461.67 Ω. At UHF, its impedance is a bit higher at the 860 MHz

end, but can be adjusted easily, by changing the width of the underpass track below

the gap. As mentioned before, the design is focussed at 960 MHz, which is the highest

end of the RFID operating frequency. At 960 MHz, the simulated impedance is 6.68 +

139.27 Ω.

7.4.6 Merging of the New UHF Dipole with the HF Coil

A dipole which provide a good matching with a UHF RFID chip at UHF is presented in

the previous section. The attention is now to merge the independently designed UHF

dipole and HF coil antenna and to obtain a resonance at HF, while maintaining a good

match at UHF.

There is no available dual-frequency RFID tag chip. It is assumed that a dual-frequency

RFID chip will not provide any input capacitance at HF. It should be pointed out

that this assumption may not be true in practice as most of the available HF chips

Page 212


5.00 10.00 15.00 20.00 25.00-12500.00

-10000.00

-7500.00

-5000.00

-2500.00

0.00

2500.00

Freq [MHz]

Figure 7.19. HF Impedance of the final version of the UHF dipole shown in Fig 7.18. (Red

curve is the resistance curve and blue curve is the reactance curve, Y-axis in Ω).

0.85 0.90 0.95 1.00 1.05 0.00

50.00

100.00

150.00

200.00

250.00

Freq [GHz]

Figure 7.20. UHF Impedance of the final version of the UHF dipole shown in Fig 7.18. (Red


have an input capacitance of 23.5 pF (Section 6.6). Nonetheless, the (capacitive) input

impedance of an RFID tag chip can be represented by a parallel capacitance and resis-

tor (Section 6.4). The additional of this parallel capacitance can be compensated easily

by reducing the overall capacitance in the HF coil of the dual-frequency antenna.

With the above stated assumption, the dual-frequency antenna itself must have a self-

resonance at HF. The individual HF antenna designed in previous section has a higher

Page 213


resonant frequency than the target resonant frequency (Fig. 7.4). Following f0 = 12π

√LC

,

where L and C are the inductance and the capacitance of the antenna respectively, ei-

ther L or C or both L and C must be increased to achieve this HF resonance. To increase

L, the number of turns can be increased or the track width can be reduced. To increase

C, the gap between tracks can be reduced.

However, before any fine-tuning was finalised, the UHF dipole must be considered as

well. As mentioned before, the UHF dipole has a gap so that it will look like an open

circuit at HF. This is not exactly true as Fig. 7.19 shows that the impedance of the UHF

dipole at HF is not near infinity.

The UHF dipole will be acting like a capacitor in parallel with the HF loop antenna

(which is inductive overall). However, the inductance and the capacitance are not high

enough to have a resonance point at 13.56 MHz, which is actually a very low frequency

to be achieved with respect to the size of the antenna.

Even with the UHF dipole attached, the four turn HF planar coil antenna is found to

have too high a self-resonance frequency as shown in Fig. 7.21. Hence the first step is to

increase both the inductance, L, and capacitance, C, of the HF antenna. The dimensions

could be modified aggressively to achieve this but there is an easier approach.

0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00-20000.00

0.00

20000.00

40000.00

60000.00

Freq [MHz]

Y 1

Figure 7.21. Self resonance frequency of HF planar coil antenna shown in Fig. 7.6. (Red


Page 214


The solution is to have a back-to-back multi-turn loop antenna on the both sides of

the FR4. Not only the length of the coil is doubled (which increases the L value), the

overlapping coil tracks on the both sides of the antenna will provide extra C to the

antenna. Via holes are used to connect the planar tracks on both sides.

Having a back-to-back multi-turn loop antenna also solves the problem in determining

the length of the transmission line, l. Since the back-to-back tracks are acting like a

capacitor, which allows UHF signals to bypass most of the inductive loop, the HF loop

antenna appears to be a short circuit at UHF. Using l equal to a quarter of a wave-

length, the short circuited HF loop antenna is transformed to an open circuit. The

design of the antenna at this stage is shown in Fig. 7.22.

Figure 7.22. Merging of HF multi-turn loop antenna with a UHF dipole. The HF loop antenna

used is the double-sided version of the HF loopantenna shown in Fig. 7.1. The UHF

dipole used is as shown in Fig. 7.18.

It can be seen that the overall dimension in the x-axis is caused by the HF antenna.

Hence the HF antenna is redesigned to reduce the overall dimension. The idea is to

have the number of loops reduced by one. To cater for the reduction in L and C caused

by the reduction of the number of loops, the coil track is made thinner, reduced from

3 mm to 2 mm. The final design is as shown in Fig. 7.23 with all the dimensions.

Fig. 7.28 shows the simulated impedance values of the dual-frequency antenna within

the UHF band. At 960 MHz, the antenna impedance is 24 + 143j Ω, which is very

Page 215


Via

Via

g

luhf

wuhf

Feed

st

wt_hfwt_uhf

wn

gn

lhf

(a)

whf

lhf

Via

Via

winner

linner

lunder

wt_uhf

(b)

Figure 7.23. Final design for dual-frequency antenna. (a) Front view, lhf = 98; luhf = 43; g =

2; wt hf = 2; wt uhf = 3; st = 6; gn = 3; wn = 16; wuhf = 37. (b) Rear view, linner

= 82, winner = 46; lunder = 38. All units in mm.

near to the intended impedance of 17 + 150j Ω. For HF, since the antenna represents

a parallel resonant circuit, the resistance is theoretically infinite at 13.56 MHz, while

the reactance is 0 Ω. The simulated impedance values of the designed dual-frequency

Page 216


4.00 6.00 8.00 10.00 12.00 14.00 16.00-10000.00

0.00

10000.00

20000.00

Freq [MHz]

Figure 7.24. Simulated impedance for the dual-frequency antenna at HF. (Red curve is the

resistance curve and blue curve is the reactance curve, Y-axis in Ω). A resonant point

is observed at approximately 13.56 MHz.

antenna at HF is shown in Fig. 7.24. A resonant point is observed at approximately

13.56 MHz. The matching in UHF impedance and the resonant point at HF shows

that the designed dual-frequency antenna has achieved the design aims shown in Sec-

tion 7.3.

7.5 Antenna Fabrication and Testing

The final design for the dual-frequency antenna (Fig. 7.23) is fabricated on a double-

sided FR4 material (Fig. 7.26) and the actual measurement is taken for both HF and

UHF operation. In the case of HF, the antenna resistance is too high at resonance to

have a proper impedance measurement using a network analyser. The unbalanced

input of the analyser also poses problems of measurement.

To demonstrate the fabricated antenna has a resonant point at HF, a transmission mea-

surement is used, where a wide-band loop antenna is used as the transmitting antenna

and the fabricated antenna is used as the receiving antenna as shown in Fig. 7.25. Ro-

tation of the loop confirms that coupling is via magnetic field. The measurement result

Page 217

7.5 Antenna Fabrication and Testing

is shown in Fig. 7.27. The exact resonance point is at 13.94 MHz, which is very near to

the designed frequency of 13.56 MHZ.

Figure 7.25. A transmission measurement using a wide-band loop antenna. The wide-band

loop antenna is placed near to the designed dual-frequency antenna with the plane of

the loop antenna in parallel with the plane of the dual-frequency antenna. When the

loop antenna is rotated 90 so that the plane of the loop antenna is perpendicular to

the plane of the dual-frequency antenna, no transmission is recorded.

Figure 7.26. Fabricated final version of the designed dual-frequency antenna.

Page 218


Figure 7.27. Transmission measurement using network analyser to locate the resonance

point at HF. Best transmission point occurs at 13.9 MHz.

Impedance measurement was carried out from 860 - 960 MHz, and compared with the

simulated results, as shown in Fig. 7.28. It was discovered that the measured resis-

tance is higher than the simulated results, but the measured reactance is lower than

the simulated results. The impedance at 960 MHz is measured to be 56.26 + 134.5 Ω.

This discrepancy is probably due to balanced load measuring problem, where the an-

tenna is a balanced load while the network analyser is unbalanced. This can be solved

by having a RF transformer or by connecting a BALUN to the antenna or employing

other means of suppressing common mode currents. However this will not be dis-

cussed further in this chapter but will be considered in the Chapter 9 of this thesis.

Nonetheless, the measurements have shown that a dual-frequency antenna has been

successfully designed and fabricated, which could be used in an RFID dual-frequency

operation.

So far we have presented a detailed design for a high frequency ratio dual-frequency

antenna. Simulated and measured results are shown to confirm its functionality at

both HF and UHF, as specified in the design aims. This shows that with proper design,

a single feed dual-antenna of very high frequency ratio can be achieved. We haven’t

focussed on the optimisation of the antenna size, which will be presented in the next

two sections.

Page 219

7.6 Miniaturisation of Dual-Frequency RFID Antenna with High Frequency Ratio

0

20

40

60

80

100

120

140

160

180

840 860 880 900 920 940 960 980

Frequency (MHz)

Impedan

ce

(Ohm

)

Simu_Re Simu_Im Meas_Re Mea_Im

Figure 7.28. The simulated and measured results for the frequency response of the fabri-

cated dual-frequency antenna within the RFID UHF band from 860 - 960 MHz.

7.6 Miniaturisation of Dual-Frequency RFID Antenna with

High Frequency Ratio

With a functioning dual-frequency antenna, the attention is now focussed on the de-

sign of a miniaturised dual-frequency Radio Frequency Identification (RFID) antenna,

which can support both the HF (13.56 MHz) and UHF bands (860 - 960 MHz). A

novel dual-frequency RFID antenna with high frequency ratio is presented in Fig. 7.23.

The dimensions of this antenna are 114 mm (l) by 98 mm (w), giving a total area of

11172 (mm)2. It has a resonant point at HF and has an impedance match to an RFID

UHF chip impedance (17-150j Ω at 915 MHz).

In designing the miniaturised version of dual-frequency antenna, some of the charac-

teristics from previous dual-frequency antenna design will be retained:

1. The new design uses a similar idea to obtain a dual-frequency antenna, which is

by designing a HF and a UHF antenna separately. These two antennas are then

merged together with a single feed. The HF antenna used is a HF coil antenna

and has a resonance point at 13.56 MHz. The UHF antenna chosen is a UHF

electric dipole with a matching network so that it has a impedance match with

the chip impedance at the operating frequency.

Page 220


2. A transmission line (with length lt in Fig. 7.41(a)) is used to link the HF and the

UHF antenna together with a single feed point, while preventing the HF antenna

from affecting the UHF antenna.

3. It is double sided. The HF coil on the top overlaps with the HF coil on the bottom

to provide enough capacitance in order to obtain a resonant point at HF.

4. A series capacitor is added at the matching network of the UHF antenna (gap pro-

vided by g in Fig. 7.41(a)). This series capacitor will have negligible impedance

at UHF but will have a very high impedance at HF. This is to prevent the UHF

antenna from shorting the HF antenna.

5. The HF antenna will provide a DC path, and this is required in certain rectifier

circuit implementations.

7.7 Improvement on Novel Dual-Frequency Antenna

In order to have a miniaturised version of a dual-frequency RFID tag antenna, some

major changes have been made on top of those retained characteristics mentioned in

Section 7.6. Also, some modifications were carried out to improve the overall perfor-

mance of a dual-frequency RFID tag antenna.

1. Relocation of UHF dipole

This improvement idea is to reposition the UHF dipole inside the HF planar coil

antenna to reduce overall size of the dual-frequency antenna. The model is as

shown in Fig. 7.29. The UHF dipole had to shrink slightly to fit.

The impedance of this antenna is as shown in Fig. 7.30 and Fig. 7.31 for HF

and UHF respectively. The HF frequency is slightly off the target frequency of

13.56 MHz, but this can be fine-tuned reasonably easily. However, the most chal-

lenging is the UHF impedance graph, where the impedances fluctuate as the fre-

quency increases. This means the antenna can only be matched to the chip at a

single frequency point only.

Page 221


Figure 7.29. A different version of a dual-frequency antenna where the dipole is located in

the inside of the HF coil. This is to reduce the overall area of the dual-frequency

antenna.

4.00 6.00 8.00 10.00 12.00 14.00 16.00-10000.00

0.00

10000.00

20000.00

Freq [MHz]

Figure 7.30. The HF impedance of the dual-frequency antenna in Fig 7.29. (Red curve is the

resistance curve and blue curve is the reactance curve, Y-axis in Ω).

2. Miniaturisation of HF coil

Since in the new design, the overall size of the dual-frequency antenna is deter-

mined by the size of the HF coil, we want a smaller HF coil. The new HF coil

is reduced from 98 mm by 65 mm to 81 mm by 58 mm, the gap is 0.5 mm and

the track width is 1 mm. The width of the transmission line is 6 mm. The rule

Page 222


860.00 880.00 900.00 920.00 940.00 960.00-50.00

0.00

50.00

100.00

150.00

200.00

250.00

Freq [MHz]

Figure 7.31. The UHF impedance of the dual-frequency antenna in Fig 7.29. (Red curve is

the resistance curve and blue curve is the reactance curve, Y-axis in Ω).

of thumb in designing an HF coil antenna is to maximise the dimension of the

interior area (winner × linner) to maximise the total inductance (including self in-

ductance and mutual inductance) of the coil. With the reduction of the size of

the HF coil, the problem of lower inductance which results in higher resonance

frequency than the intended frequency will be encountered. This problem can be

solved by careful design and by introducing an additional loop into the HF coil

as shown in Fig. 7.32.

Figure 7.32. A miniaturised HF coil antenna. With reduced track width and spacing.

3. Miniaturisation of UHF dipole

Page 223


With the size reduction of the HF coil antenna, the UHF dipole antenna can be

reduced as shown in Fig. 7.33. Basically, the size of the UHF antenna together

with its matching network is designed to have a thinner track width, shorter

length along x-axis, but with a longer bent side.

Figure 7.33. A miniaturised UHF dipole. With thinner tracks and longer side bending.

The miniaturised HF coil antenna (Fig. 7.32) and UHF dipole (Fig. 7.33) are then

merged together to obtain a new and miniaturised dual-frequency antenna as

shown in Fig. 7.34.

4. HF not affecting UHF over a broader bandwidth

Initially, it was believed that the HF antenna would behave like a short circuit at

the UHF band as the overlapping HF coil provides a short circuit path for UHF

signal (Refer Section 7.4.3). Hence, it was decided to use a transmission line of

length λ4 to transform this short circuit into an open circuit, so that the HF coil

would not affect the operation of the UHF dipole.

However, it was later discovered that the above statement is only true for a single

frequency point. The HF coil antenna is not exactly a short circuit for the entire

band of UHF as shown in Fig. 7.35(a). If a λ4 transmission line of any characteristic

impedance is simply introduced into the structure, the HF antenna will still affect

the UHF antenna over a large portion of UHF band.

Page 224


Figure 7.34. A dual-frequency antenna with miniaturised HF coil and UHF dipole. Formed

using the miniaturised HF coil antenna shown in Fig. 7.32 and the miniaturised UHF

dipole shown in Fig. 7.33

Hence, the idea is to design a transmission line with high characteristic impedance,

Z0. The type of transmission line chosen consists of coplanar strips (CPS). CPS

has a theoretical physical limit on the lowest and highest Z0 of 45 and 280 Ω re-

spectively when εr is 10 [86]. The highest Z0 limit can be increased by reducing

the εr value.

The Z0 of the CPS can be increased by changing wt and st (Refer Fig. 7.41(a)). If

the CPS has a very high Z0, the impedance of the designed antenna can be re-

normalised in Smith Chart (by default normalised by Z0 of 50 Ω) to the left hand

region of the Smith Chart. Then, this impedance can be transformed to the right

hand region using a λ4 long CSP.

In mathematical terms, the reflection coefficient, Γ of an antenna is given by:

Γ =ZL − Z0

ZL + Z0(7.2)

where the characteristic impedance (or normalising impedance), Z0, is purely

resistive. Rearranging (7.2) will give us:

Page 225


1.0 1.00.0

0

10

20

30

40

50

6070

8090100110

120

130

140

150

160

170

180

-170

-160

-150

-140

-130

-120-110

-100 -90 -80-70

-60

-50

-40

-30

-20

-10

5.002.001.000.500.200.000.000.000.000.000.000.000.000.00

5.00

-5.00

2.00

-2.00

1.00

-1.00

0.50

-0.50

0.20

-0.20

0.0

1

2

(a)

1.0 1.00.0

0

10

20

30

40

50

6070

8090100110

120

130

140

150

160

170

180

-170

-160

-150

-140

-130

-120-110

-100 -90 -80-70

-60

-50

-40

-30

-20

-10

5.002.001.000.500.200.000.000.000.000.000.000.000.000.00

5.00

-5.00

2.00

-2.00

1.00

-1.00

0.50

-0.50

0.20

-0.20

0.0

1

(b)

Figure 7.35. Impedance transformation in Smith Chart where all the traces cover 860 to

960 MHz. (a) Trace 2 is the reflection coefficient of the impedance of the antenna

normalised to 50 Ω while Trace 1 is the reflection coefficient of the impedance of the

antenna re-normalised to 290 Ω. (b) Trace of the reflection coefficient when Trace 1 in

(a) is transformed using a λ4 transmission line. Note that the length of the transmission

line corresponds to λ4 of a particular frequency only (not the entire frequency band).

Hence (b) is not exactly (only similar to) Trace 1 of (a) rotated by 180.

Page 226


ZL = Z01 + Γ1 − Γ

(7.3)

And with a new characteristic impedance (of the CPS), the new Γ can be com-

puted using:

Γnew =Z0

1+Γ1−Γ − Znew

0

Z01+Γ1−Γ + Znew

0(7.4)

The effect is shown in Fig. 7.35(a), and with the impedance transformation of theλ4 CPS for the designed frequency (approximately λ

4 for other frequencies within

the band of operation), the HF coil impedance is plotted as Fig. 7.35(b). It can be

seen that the HF coil behaves like a high impedance for most of the UHF band of

interest.

The final design of the transmission line linking the HF and the UHF antenna

together has the following design values (Refer Fig. 7.41(a)): st = 6 mm, wt =

0.8 mm and εr = 4.4, resulting a Znew0 ≈ 290 Ω. The effective dielectric constant

value, εe f f is approximately 2.12.

Hence, the length of the λ4 CSP is:

lt =λ

4 ×√εr

(7.5)

At 920 MHz, λ is approximately 0.326 m. Hence lt is approximately 0.056 m or

56 mm.

5. The coupling effect between UHF and HF antenna

As mentioned before, a UHF electric dipole antenna with matching impedance to

the RFID tag chip was designed and simulated. A transmission line was used to

separate the HF and UHF antennas. In the previous dual-frequency design, the

UHF dipole antenna is located beside the HF coil. Although the overall size is in-

creased, the coupling between the HF and UHF antennas is negligible. In the new

design, it is decided to move the UHF dipole inside the HF coil to reduce over-

all size. However, the coupling between the antennas changes the UHF dipole

Page 227


impedance quite significantly and the UHF dipole no longer has an impedance

match with the UHF chip at UHF. A fine tuning was carried out by re-adjusting

g, luhf and lunder.

6. Reduction of HF antenna size

To fit both the transmission line and the UHF antenna within the HF coil an-

tenna, the HF coil antenna is rotated 90 degree from the previous design (com-

pare Fig. 7.18 and Fig. 7.36). With the HF coil rotated, the simulation results are

as shown in Fig. 7.37 and Fig. 7.38 for the antenna impedance in HF and UHF

respectively. It can be seen that it is still functioning well as a dual-frequency

antenna.

Figure 7.36. Dual-frequency antenna with rotated HF coil. The rotation of HF coil does not

change much on the characteristics of the dual-frequency antenna.

However, the antenna in Fig. 7.36 does not contribute to the aim of minimising

the size of the antenna. Hence, ultimately, the UHF antenna has to be minimised

and relocated inside the HF coil antenna as discussed before. The impact on the

size limitation of the UHF dipole antenna is significant. The wuhf is shortened

from 95 mm to 44 mm. Although this dipole antenna can be designed to match

in impedance with the RFID chip, its simulated directivity drops from 2.9 to 2.2.

This is a trade-off between size and performance.

Page 228


5.00 10.00 15.00 20.00 25.00-20000.00

-10000.00

0.00

10000.00

20000.00

30000.00

Freq [MHz]

1

Figure 7.37. Simulated results at HF for antenna in Fig. 7.36. (Red curve is the resistance

curve and blue curve is the reactance curve, Y-axis in Ω). Point 1 indicates a resonant

point at 12.30 MHz.

860.00 880.00 900.00 920.00 940.00 960.00 0.00

50.00

100.00

150.00

200.00

Freq [MHz]

Figure 7.38. Simulated results at UHF for antenna in Fig. 7.36. (Red curve is the resistance

curve and blue curve is the reactance curve, Y-axis in Ω).

Note that the six improvements presented in this section are not subsequence improve-

ments. For example, the reduction of HF antenna size (improvement 6) does not take

place after all the first five improvements. All these improvements complement each

other and contribute to the final design of the miniaturised dual-frequency RFID tag

antenna. The complete blue print of the final design can be found Fig. 7.41.

Page 229

7.8 Testing on Miniaturised Antenna

A B

Figure 7.39. Figure to illustrate the position of RFID chip attached to the antenna. In

the initial testing both HF and UHF chips are attached to feed point A. In the final

prototype, UHF chip is located at point A while HF chip is located at point B. The

impedance of the transmission line is approximately 290 Ω.

7.8 Testing on Miniaturised Antenna

The prototype of the designed dual-frequency antenna (Fig. 7.41) was tested using the

methods described in Section 7.5, and found to be resonating at 13.56 MHz while hav-

ing reasonable impedance match with a generic UHF RFID tag chip at UHF. Hence the

novel dual-frequency RFID tag antenna shown in Fig. 7.23 has been miniaturised suc-

cessfully while its HF and UHF operational characteristics are preserved in achieving

the design aims presented in Section 7.3.

The next testing procedure is to measure the read range performance of the minia-

turised dual-frequency antenna. The biggest challenge in real life testing for a dual-

frequency RFID tag is that there is no single tag chip available that would work on

both HF and UHF. It is the design aim to have a single feed system. However, for the

purpose of testing, an UHF RFID C1G1 chip and a Tag Talk First HF chip (72128 TTF

chip on smart card module with negligible input capacitance) are used instead.

Referring to Fig. 7.39, both the HF (72128 TTF) and UHF chip (Alien C1G1) are attached

at position A. In theory, the read range of HF operation is limited by the size of the

reader antenna [95]. A quick measurement shows that this dual-frequency RFID tag

offers a maximum possible read range at HF when using a square loop HF reader

antenna with edge-to-edge length of approximately 0.3 m.

Page 230


However, the HF chip affects the UHF chip significantly and a poor read range of only

0.3 m is observed at UHF. If the HF chip is moved to the B position in Fig. 7.39, the

HF chip will be transformed to a very high impedance at UHF by the transmission

line. Hence the UHF chip will perform normally and a read range of more than 2 m is

observed. A quick check again shows that this dual-frequency RFID tag still offers a

maximum possible read range at HF.

Although the final prototype dual-frequency RFID tag has two chips and has two feed

points, it is believed that this prototype will work with equivalent efficiency in both

HF and UHF bands when a single dual-frequency chip is used on a single feed.

To compare this miniaturised dual-frequency RFID tag with commercial tags, a com-

mercial HF RFID tag and a commercial UHF RFID tag are placed side by side and

a read range measurement in both the HF and UHF bands is carried out. It is dis-

covered that the read range performances are comparable when the tags are alone or

placed near to each other. The conclusion is that, the prototype dual-frequency an-

tenna antenna is superior to the commercial tags because it can handle operations in

both frequency bands with only one feeding point. More discussions on the dual feed

point dual-frequency antenna can be found in Section 8.5.

7.9 Final Design of Miniaturised Dual-Frequency Antenna

The final design of the miniaturised dual-frequency antenna is as shown in Fig. 7.40.

In short, the new design is to relocate the UHF electric dipole antenna inside the HF

coil antenna to reduce the overall size as shown in Fig. 7.41. The total dimension is

81 mm (lhf) by 58 mm (whf), giving a total area of 4698 (mm)2.

7.10 Conclusion

A novel design for a dual-frequency RFID antenna with high frequency ratio was pre-

sented by the method of merging an HF antenna together with a UHF antenna. This

Page 231

7.10 Conclusion

Figure 7.40. A miniaturised dual-frequency antenna. With HF and UHF RFID chips attached.

Via

Via

Feed A

l

w

l

w

t

t

luhf

wuhf

g

hf

hf

ts

Feed B

(a) Front view

l

innerw

inner

Via

Via

lunder

(b) Back view

Figure 7.41. Final antenna design with design parameters. The final prototype has the following

design values in mm: lhf = 81; whf = 58; linner = 70; winner = 48; luhf = 37; wuhf =

44; g = 3; wt = 0.8; st = 6; lunder = 9.

antenna works well in both the HF and UHF bands for RFID operation with a sin-

gle feed. A functioning dual-frequency RFID tag would prove that having a dual-

frequency RFID system is feasible. Although no dual-frequency RFID chip is available

at the moment, it is hoped that this successful design of a compact dual-frequency

Page 232


RFID antenna will catalyse the development of a dual-frequency RFID chip. A minia-

turised version was also presented to demonstrate the potential of this dual-frequency

antenna.

Page 233

Chapter 8

Alternative Dual-frequencyAntenna Designs

THISchapter complements the research of a novel dual-frequency

antenna of high frequency ratio presented in previous chapter, by

exploring alternative methods in creating a dual-frequency an-

tenna of similar characteristics.

Page 235

8.1 Introduction

8.1 Introduction

As described in Chapter 7, three methodologies were brainstormed to create a single

feed point dual-frequency RFID antenna which operates in both HF (13.56 MHz) and

UHF (860 - 960 MHz). These three methodologies are:

1. Merging an HF antenna and a UHF antenna.

2. Making an HF antenna that operates at UHF.

3. Making a UHF antenna that operates at HF.

Chapter 7 has presented a fully functional dual-frequency RFID tag antenna with its

miniaturised version, developed based on the first methodology, where individual HF

and UHF antennas are designed separately, and combined in a way where the interfer-

ence between each other are minimised even with only a single feed point.

This chapter will cover the alternative methods of merging an HF antenna and a UHF

antenna, and also the remaining two brainstormed methodologies, which are (1) in

Section 8.3, making an HF antenna that operates at UHF, and (2) in Section 8.4, making

a UHF antenna that operates at HF.

This chapter will also investigate the case where two independent HF and UHF RFID

tag are located near to each other, to create a dual-frequency tag but with two isolated

antenna and two independent chips. This is presented in Section 8.5, followed by in

Section 8.6 the comparisons of all methods presented in this thesis to create a novel

dual-frequency antenna, and in Section 8.7 at the end of this chapter by conclusions.

8.2 Alternate Methods in Merging an HF Antenna and

a UHF Antenna

Two methods different from the approach shown in Chapter 7 are investigated for

merging an HF antenna and an UHF antenna, to create a dual-frequency antenna. The

Page 236

Chapter 8 Alternative Dual-frequency Antenna Designs

first approach is to change the position of the feed point while the second approach

uses different UHF antennas (the UHF antenna used in the design in Chapter 7 is a

UHF dipole antenna) for the merging.

8.2.1 Different Feed Point Location

As discussed in the end of Section 7.4.3 and shown in Fig. 7.9, there are two possible

selections of feed point position, at position A or position B. Also the arguments pre-

sented in that section concluded that a choice of position A may make the fine-tuning

process of a dual-frequency antenna difficult as compared to the choice of placing the

feed point at position B. In this section, the results obtained in the case where the feed

point is repositioned to location A are presented and from the following results it is

concluded that a change in the feed location from location B to location A is not feasi-

ble in creating a dual-frequency antenna.

lA B

Figure 8.1. Feed point repositioning of the dual-frequency antenna. Feed point relocated from

position B to position A.

The dual-frequency antenna of interest is shown in Fig. 7.23. This antenna was chosen

as it has been proven in Chapter 7 as a fully functional dual-frequency antenna at both

HF and UHF bands. In the original form, this antenna has a feed point at position

B. The feed point has been relocated to position A for the analysis presented in this

section.

Simulation was carried out using Ansoft HFSS. The simulation results are as shown in

Fig. 8.2 and Fig. 8.3, showing the impedances of the dual-frequency antenna at HF and

UHF respectively with the new feed point.

Page 237

8.2 Alternate Methods in Merging an HF Antenna and a UHF Antenna

4.00 6.00 8.00 10.00 12.00 14.00 16.00-10000.00

0.00

10000.00

20000.00

Freq [MHz]

Figure 8.2. Impedance of dual-frequency antenna with feed point at position A at HF. (Red

curve is the resistance curve and blue curve is the reactance curve, Y-axis in Ω). It can

be seen that a change in feed location does not bring much effect to the impedance of

the antenna at HF.

860.00 880.00 900.00 920.00 940.00 960.000.00

50.00

100.00

150.00

200.00

250.00

300.00

Freq [MHz]

Figure 8.3. Impedance of dual-frequency antenna with feed point at location A at UHF.


As discussed in Section 7.4.3, when the feed point is chosen at location B, the UHF

dipole will contribute useful additional capacitance to the HF antenna when the merged

antenna is operating at HF. With a change in feed point from position B to A, the

changes in reactance at HF would be negligible. This is proven by Fig. 8.2 where the

resonant point is just below 13.5 MHz, very near to the original 13.56 MHz (Fig. 7.24).

Page 238


Note that though there only a negligible change in the resultant reactance, the contri-

bution of the transmission line to the total reactance is in a different form when the

position of feed point was changed from position B to position A. When the feed point

is at position B, the transmission line is contributing parallel capacitance while when

the feed point is at position A, the transmission line is contributing series inductance.

However, the change of feed point brings significant change to the impedance of the

antenna of interest (the dual-frequency antenna in Fig. 7.9 which consists of a UHF

dipole and an HF coil) at UHF. With that change of feed point, the λ4 (UHF) transmis-

sion line is not functioning as intended, where it is supposed to transform the HF coil

impedance into a relatively high impedance at UHF. Now the impedance at UHF of the

UHF dipole is transformed to an undesirable value while the HF coil may be acting as

the main radiating body at UHF.

A quick fix might be thought to be is to remove the λ4 (UHF) transmission line com-

pletely so that the impedance of the UHF antenna is not transformed. However, there

is then no way to avoid the impedance of the HF antenna from disturbing the input

impedance at the feed point.

8.2.2 Different UHF Antenna Types

Section 7.4.1 presented the individually designed HF antenna and UHF antenna, before

they are merged together using a transmission line to have a single feed point. The

chosen HF antenna is a multi-turn planar coil antenna, while the chosen UHF antenna

is an electric dipole. This section will explore the alternative antennas to be used as the

UHF antenna.

1. UHF Patch Antenna

A simple UHF patch antenna is as shown in Fig. 8.4. The patch has a width of

99 mm and length 77 mm with the ground plane width and length of 120 mm and

100 mm respectively. The dielectric is a 1.6 mm FR4 with dielectric permittivity

of 4.4.

Page 239

8.2 Alternate Methods in Merging an HF Antenna and a UHF Antenna

Figure 8.4. A UHF patch antenna.

From simulation, the best match point occurs at approximately 880 MHz with

an input impedance of 15 + j70 Ω. It is not a perfect match with any of the UHF

chips shown in Section 6.6. However, from experience, with minimal fine-tuning,

this antenna is definitely an acceptable candidate to be used as a UHF RFID tag

antenna.

Also from simulation, this UHF patch antenna is capacitive at HF. This is a de-

sirable result as discussed in Section 7.4.6, where the UHF antenna is responsible

to provide extra capacitance to the HF coil so a resonance at 13.56 MHz can be

achieved.

Hence, an alternative way to create a dual-frequency antenna is to swap the UHF

electric dipole with a UHF patch antenna. The short-coming of this method is

that the UHF patch antenna cannot be positioned within a HF coil antenna as

compared to the case where an electric dipole can be relocated within a HF coil

antenna (Fig. 7.40). This will unavoidably increase the overall size of the dual-

frequency antenna as the total area is mostly contributed by the area occupied by

both the UHF patch antenna and the HF coil antenna.

2. UHF Loop Antenna

Page 240


A UHF RFID tag antenna in the form of loop antenna has been designed and

presented in Section 6.11.3. This loop antenna has a read range of under 1.5 m.

With the inclusion of matching network, this UHF loop antenna is capacitive at

HF. Hence, it can be used to be merged with an HF coil antenna to create a dual-

frequency antenna.

In short, any UHF antenna with matching impedance at UHF with respect to a UHF

RFID chip, and is capacitive at HF, can be used as the UHF antenna to be merged

with an HF coil antenna to form a dual-frequency antenna, using the merging tech-

nique where a HF antenna and a UHF antenna are joined by a λ4 transmission line

(Section 7.4.6). If a UHF antenna is not capacitive at HF, a series capacitor can be added

to achieve this.

8.3 HF Antenna Acting as UHF Antenna at UHF

This section investigates the performance of an HF antenna at UHF. The aim is to de-

sign an HF antenna which operates at HF at the beginning, and with proper tuning

later, it is hoped that this same HF antenna will also have a good match with a UHF

chip. If this is the case, this HF antenna will be acting as a UHF antenna at UHF, and

this HF antenna is in fact a fully functional dual-frequency antenna.

The antenna of interest is an un-tuned HF antenna (Fig. 8.5), and this antenna is de-

signed to resonate in the HF range between 20 MHz to 30 MHz.

The simulated impedance of this un-tuned HF antenna is as shown in Fig. 8.6. From the

figure, it can be seen that the resonance frequency is approximately 25.4 MHz, which is

higher than the desired frequency. However, the focus here is not to make this simple

HF antenna to resonate at 13.56 MHz, though, it can be done by adjusting the parallel

capacitance and choosing the right HF chip as shown in Section 6.10.4.

As mentioned before, the aim in this section is to examine whether it is feasible to

have an HF antenna operating at UHF. The simulation results for the same HF antenna

(Fig 8.5) at UHF are as shown in Fig. 8.7.

Page 241


Figure 8.5. The HF antenna to be fine-tuned to operate at UHF. The feed point of the antenna

is represented by a small rectangular box. The red colour track is the top coil while

the gray colour track is the overlapping coil on the other side of the dielectric material

(FR4). This HF coil antenna is the early version of HF antenna designed for pig tagging

purposes (Section 6.10.4).

15.00 17.50 20.00 22.50 25.00 27.50 30.00-10000.00

0.00

10000.00

20000.00

Freq [MHz]

Figure 8.6. Simulated impedance of the HF antenna shown in Fig. 8.5 at HF. (Red curve

is the resistance curve and blue curve is the reactance curve, Y-axis in Ω). Resonant

point at 25.4 MHz.

The impedance around 800 MHz to 850 MHz is almost the intended matching impedance

with respect to a RFID tag chip (Best match point at approximately 828 MHz). With

Page 242


750.00 770.00 790.00 810.00 830.00 850.00 0.00

50.00

100.00

150.00

200.00

Freq [MHz]

1

2

3

4

Figure 8.7. Simulated impedance of the HF antenna shown in Fig. 8.5 at UHF. (Red curve is

the resistance curve and blue curve is the reactance curve, Y-axis in Ω). Points 1 and

2 indicates the best match point at approximately 828 MHz.

proper tuning, it is confirmed through simulation that this HF antenna can indeed

match in impedance to a UHF chip.

Despite the fact that an HF antenna can be matched to a UHF RFID chip, there remain

a few problems that need to be looked into. Firstly, this antenna is difficult to be fine-

tuned to operate at both HF and UHF. For example, from experience, if the antenna is

tuned to resonate at HF, it will become unmatched in UHF, and vice versa.

Secondly, a normal HF coil has a multi-turn structure to provide sufficient inductances

to obtain resonance at 13.56 MHz. The multi-turn structure is relatively long in terms

of UHF wavelength, and hence even with the impedance matched, it is suspected that

it will be too lossy to be used as a UHF radiator. Evidence on this matter will be

presented in the next section.

Page 243


8.3.1 Loss in Antenna

The loss of an antenna can be estimated using the simulated directivity and gain. In

simple terms, gain is the directivity of an antenna, taking into consideration the dissi-

pative loss of an antenna (a thorough discussion of directivity and gain can be found

in Section 6.3).

For the antenna shown in Fig. 8.5 with the feed point indicated by a rectangle in the

same figure, the peak directivity obtained through simulation is 3.5476 but its peak

gain is 0.00046716. This proves that this antenna is a very poor radiator. The following

discussion explains why.

The plot of the current density at frequency 910 MHz is included here as Fig. 8.8. As in

Fig. 8.5, the feed point is represented by a small rectangle in the figure. This antenna

(from Fig. 8.5) has two sides of multi-turn loops. The top side loops are not obscured

by the bottom side loops, however, the bottom side loops (which are gray) are hardly

seen in Fig. 8.8 as they overlap with the top side loops. For the top side loops only, the

current density is shown at a particular point in time using the colours indicated in the

legend.

Figure 8.8. The structure of the HF antenna shown in Fig. 8.5 with illustration of surface

current density on the top side. The feed point is represented by a small rectangular

box.

Page 244


Upon a careful inspection, it can be seen that there is a standing wave pattern in the

plot of current density. We conclude that this antenna has too long a track for operation

at 910 MHz. Current changes phase along the multi-turn coil, with approximately a

phase change every turn. Radiation from currents flowing in adjacent turns will then

tend to cancel.

Though the current density shown in Fig. 8.8 is a just snapshot of the current density,

this figure shows that the radiation properties of this antenna are hard to control at the

frequency of interest.

8.3.2 Antenna Measurement Results

As discussed in previous section, it is highly suspected that insufficient power will

reach the feed point of the antenna when the antenna is receiving.

To confirm the claim that the HF antenna is too lossy to be used as a UHF antenna,

the HF antenna shown in Fig. 8.5 is fabricated with a UHF chip is attached to it. A

quick measurement test was devised. In this test, a small wide-band loop antenna

(with diameter of approximately 20 mm), was connected to a network analyser. With

the network analyser measuring the reflection at the port connected with that loop

antenna, the HF antenna with a UHF chip was positioned near that loop antenna. The

reflection measurement provided a good estimation on how efficiently the HF antenna

was acting as a UHF absorber. The results is shown in Fig. 8.9. A measurement of

< -3 dB reflection (with output power of network analyser set to -20 dBm) shows that

sufficient energy has been absorbed by the antenna for a normal UHF tag operation.

However, when the tag was positioned within the interrogation zone of a UHF reader,

the tag cannot be detected. The distance between the tag and the reader antenna was

reduced slowly towards zero, but no tag detection was recorded by the reader. As

mentioned before, this is most likely caused by the lossy nature of the HF antenna at

UHF and all the energy absorbed by the antenna has been dissipated in the coil, rather

than transferred into the UHF chip.

Page 245

8.4 UHF Antenna Acting as an HF Antenna at HF

Figure 8.9. Measurement result of the HF antenna shown in Fig. 8.5 at UHF. The measure-

ment result is identical to the frequency response of a functioning UHF tag.

This concludes the experiment to make use of an HF antenna as a UHF antenna at UHF.

The truth seems to be an HF antenna, which has a size feasible for RFID application

and to resonate at 13.56 MHz, will be lossy or have an unsuitable current distribution

(not reaching the feed point) at UHF. Even though it can be fine-tuned to match in

impedance to a UHF chip, the losses in the coil or the unsuitable current distribution

prevent sufficient power from reaching the UHF chip.

8.4 UHF Antenna Acting as an HF Antenna at HF

This is the least explored methodology. The reason is that under normal operating

conditions a UHF antenna would not have high net inductance value at HF. Two UHF

antennas are considered, namely the UHF dipole and the UHF patch antenna. From

previous discussions, it has already been shown that all these two UHF antennas are

capacitive at HF. Hence, it is concluded that without an additional matching circuit,

a common UHF RFID tag antenna would not be suitable to be used as an RFID HF

tag antenna. The inclusion of matching network is not a solution as it will de-tune the

UHF antenna and the matching between the UHF antenna and the UHF chip will be

Page 246


affected. Also, if a coil UHF antenna (with few turns) is used, a matching over too large

an impedance ratio is required, if this coil antenna is to be tuned to operate at HF.

8.5 Multi-Feed Point Antenna

Although it is the aim to have a single feed dual-frequency antenna, it is of great in-

terest to examine what will happen if one HF and one UHF antenna are placed near to

each other, with two separate chips.

The advantages of having two separate antennas are:

1. HF and UHF tags are readily available. No research or development is required.

2. Modification of the chips, which is necessary for a single feed dual-frequency

antenna, is avoided.

3. Even if redesign of HF and UHF antennas is involved, it is not as complex as the

design of a single dual-frequency antenna.

On the other hand, the disadvantages of this configuration are:

1. The need to program the HF tag and UHF tag to have a same EPC, or a database

is required to link the HF and UHF EPC together.

2. Two chips are required as compared to one needed in a single feed dual-frequency

antenna.

3. The configuration may attract a higher price, as it is basically two tags as com-

pared to one tag based on one single feed dual-frequency antenna.

8.5.1 Stacking UHF and HF Antennas

The idea is to stack a UHF antenna on top of a HF antenna, reducing the horizontal area

needed, but increasing the vertical volume at the same time. This design is based on the

HF pig tag presented in Section 6.10.4 and the UHF pig tag discussed in Section 6.11.4.

The reasons these two antennas are chosen are:

Page 247


• The sizes are almost the same.

Since both the HF and UHF RFID tags designed for pig tagging use the same

encapsulation casing (Fig. 6.29), they are of similar size.

• There is an opening in the middle of the chosen UHF antenna.

An HF antenna will only work when there exists a magnetic field which couples

through the HF coil. To make this possible, the chosen UHF antenna must have

an opening in the middle.

• These antennas can be readily encapsulated using the casing shown in Fig. 6.29.

It is measured that the maximum spacing between the two antennas is 2 mm so

that they can be fitted into the casing.

Fig. 8.10 shows the suggested configuration. The separation distance between the HF

and UHF RFID tags, t, is controlled using small pieces of thin paper.

t

UHF

HF

Figure 8.10. Stacking of HF and UHF RFID tags. The HF and UHF RFID tags are separated

a distance, t, away. Pieces of thin paper are inserted or removed between the two

antennas to control the spacing.

For the HF operation, as long as the UHF antenna is not touching the HF antenna (no

electric conduction), the HF tag can be detected by a HF reader. However, this is not

found to be the case for the UHF tag where the HF tag has a great impact on the read

range.

Read range measurements were carried out while the spacing between these two an-

tenna was varied and the results are recorded in Table 8.1.

When the two tags are too close to each other, the UHF tag was de-tuned and could not

be detected. To re-tune the tag using a network analyser is unexpectedly complex as

Page 248


Table 8.1. Measurement results for the UHF read range of the dual-frequency antenna using

the stacking method.

Separation (mm) Read Range (m)

∞ 0.75

10 0.38

9 0.35

5 0.30

the HF tag has some resonant points in the UHF band of interest, as shown in Fig. 8.11.

When the HF tag and UHF tag are placed near to each other, the response looks like

Fig. 8.12, where it is impossible to distinguish which is the UHF response. Fig. 8.13

presents the frequency response when the UHF tag is 5 mm away from HF tag, and

can be interrogated by a reader.

Figure 8.11. Frequency response of the HF tag at UHF without the presence of any UHF

tag. It can be seen that there are several dips in the UHF band caused by an HF tag.

Page 249


Figure 8.12. Frequency response of the UHF tag at UHF when placed near to HF tag. The

distance between the HF and UHF tags is approximately 2 mm. As it can be seen on

the above measurement results, it is very difficult to distinguish the frequency response

of the UHF tag from the HF tag for fine-tuning purposes.

Figure 8.13. Frequency response of the UHF tag at UHF when the tag is able to be read

by an interrogator. The separation between the HF and UHF tag is 5 mm.

Page 250


The results showed that a separation between the UHF and the HF antenna is required

for the UHF tag to operate. Nonetheless, the method of stacking can be used to create

a dual feed point dual-frequency RFID tag.

8.5.2 Side-by-side HF and UHF Antenna

In this section, two RFID tags (one HF and one UHF) are placed side-by-side. A com-

mercial UHF tag is often an electric dipole with a matching circuit and is long and thin.

A commercial HF tag on the other hand is always a multi-turn coil antenna. The area in

the middle of the multi-turn coils is a major factor in controlling the total inductance.

A quick test proved that above the HF and UHF tags can be read by HF and UHF RFID

readers as long as the two tags do not overlap each other.

8.6 Comparison Between Different Methodologies

This section compares all the methods presented in this thesis which were used to

develop a dual-frequency RFID tag antenna. Some methods yield better results than

the others.

Table 8.2 summarises the observations on simulation and measurement results. The

best results are from the merging of a UHF antenna with an HF antenna. The dual-

frequency antenna produced using this method, with the UHF antenna being a UHF

dipole and the HF antenna being an HF coil, was presented in Chapter 7.

The choice of the antenna types in using the merging method in producing a dual-

frequency antenna is important. for example, merging an HF coil antenna with a UHF

patch antenna yields the same satisfactory performance, but requires a bigger area

(Section. 8.2.2) while merging an HF coil antenna with a UHF coil antenna produces a

satisfactory dual-frequency antenna (Section 8.2.2).

Page 251

8.6 Comparison Between Different MethodologiesTable

8.2

.Com

par

ison

bet

wee

ndiff

eren

tm

ethodolo

gie

sin

crea

ting

adual-freq

uen

cyRFID

tag

ante

nna.

The

Ref

eren

ceco

lum

nsh

ows

the

loca

tion

inth

isth

esis

wher

eth

ediscu

ssio

nof

diff

eren

tm

ethodol

ogie

sca

nbe

found.

The

Fee

dPoi

nt

Tot

alco

lum

nsh

ows

the

num

ber

of

feed

poi

nt

required

while

the

Fee

dPoi

nt

Loca

tion

isth

epos

itio

nof

feed

poi

nt

with

resp

ect

toth

eco

nfigu

ration

show

nin

Fig

.7.

9.T

he

HF

and

UH

Fco

lum

ns

indic

ate

whet

her

ornot

the

des

igned

dual

-fre

quen

cyan

tenna

isab

leto

oper

ate

inH

Fan

dU

HF

ban

ds

resp

ective

ly.

The

aste

risk

sign

,∗ ,

indic

ates

ades

igned

dual

-fre

quen

cyan

tenna

isei

ther

wor

king

inth

eH

For

the

UH

Fban

d,but

not

inbot

hban

ds

as

inte

nded

.

Met

hodo

logi

esR

efer

ence

Feed

Poin

tH

FU

HF

Tota

lLo

cati

on

Mer

ging

(UH

Fdi

pole

and

HF

coil)

Cha

pter

71

B

Mer

ging

(UH

Fpa

tch

and

HF

coil)

Sect

ion.

8.2.

21

B

Mer

ging

(UH

Fco

ilan

dH

Fco

il)Se

ctio

n8.

2.2

1B

Mer

ging

(UH

Fco

ilan

dH

Fco

il)Se

ctio

n8.

2.1

1A

∗

∗

HF

asU

HF

ante

nna

atU

HF

Sect

ion

8.3

1N

/A

×

UH

Fas

HF

ante

nna

atH

FSe

ctio

n8.

41

N/A

×

Stac

king

ofH

Fan

dU

HF

Sect

ion

8.5.

12

N/A

Side

-by-

side

HF

and

UH

FSe

ctio

n8.

5.2

2N

/A

Page 252


Changing the feed point location from position B to position A (Fig. 8.1) is not an option

as the HF antenna and the UHF antenna will interfere each other (Section 8.2.1). The

end result is either the dual-frequency antenna works at the HF band only or at the

UHF band only.

It is also proven that an HF antenna will not be an efficient radiator at UHF (Sec-

tion 8.3), and a UHF antenna will not be able to provide sufficient inductance for HF

chip at HF (Section 8.4).

Using a multi-feed structure is a viable option (Section 8.5.1 and Section 8.5.2). The

only short-coming is that two chips are needed instead of one. However, it is the design

aim to have just a single feed dual-frequency antenna to reduce the cost. Hence, the

best design as recommended by this thesis is the merging of UHF dipole and HF coil

antenna to form a single feed high frequency ratio dual-frequency RFID tag antenna.

8.7 Conclusion

This chapter has offered detailed discussion on all the methods brainstormed in cre-

ating a dual-frequency antenna apart from the most successful method which has al-

ready been presented in a previous chapter. A comprehensive comparison between the

strengths and weaknesses of each method is also included in this chapter. This shows

that there are some flexibilities in designing dual-frequency antennas depending on

the application. This is important as a wide deployment may require various forms of

dual-frequency antenna.

Page 253

Chapter 9

Measurement Techniques

THE performance of a passive Radio Frequency Identification

(RFID) tag depends on the antenna and the chip efficiency of the

tag, and also depends on the matching between the antenna and

the chip. The design of the tag antenna often involves the measurement

of its impedance. This chapter investigates the RF cable effect on the mea-

surement of RFID tag antenna impedance. Measurement results obtained

through several common methods are compared with the simulated results.

Ways of minimising measurement error are discussed and an analysis of the

effect of mismatch between the tag antenna and the chip is presented.

Page 255

9.1 Introduction

9.1 Introduction

Various RFID tag antenna designs for HF and UHF RFID tags have been presented

from Chapter 6 to Chapter 8, with simulation and measurement results presented. The

simulation of an RFID tag antenna was carried out using Ansoft HFSS as presented in

Section 6.9.

During simulation, fine-tuning can be carried out to adjust antenna performance pa-

rameters, such as the operating frequency, bandwidth and impedance, to the desired

values. Most of the time, with the restriction of antenna size and cost, the design-

ing stage often involves trade-offs between some of the antenna performance param-

eters. The designed antenna is then fabricated. The biggest challenge is to discover

how a prototype tag antenna can be tested to confirm that it functions as designed,

before large scale production is carried out. Hence, this chapter aims to investigate

the challenge of measuring a small RFID tag antenna using several different methods.

Suggestions are presented to improve the measurement results to reflect the actual

characteristic of an RFID tag antenna.

Though this chapter discusses the measurement techniques devised to obtain the in-

put impedance of the prototype RFID tag antennas, the focus of this chapter is on the

measurement techniques for UHF RFID tag antennas. The measurement of HF RFID

tag antenna is relatively simple for the following reasons:

• HF (13.56 MHz) is relatively low in frequency as compared to UHF (860 - 960 MHz).

It is a known fact that high frequency measurements are more challenging than

low frequency measurements.

• An HF tag antenna is normally low in resistance (at negligible level) and is induc-

tive (a multi-turn coil antenna). A direct measurement using a network analyser

will give a good approximation of the resultant reactance value at 13.56 MHz.

Alternatively, a known value capacitor can be attached across the feed points of

the HF antenna, and by exciting the HF antenna while monitoring the network

analyser, the resonant frequency and the resultant inductance value of the HF

antenna can be obtained. Though, the stated methods are not ideal, as there are

Page 256

Chapter 9 Measurement Techniques

several other problems such as the problem of a balanced load connected to un-

balanced measuring device, experience shows that the impact is not severe, and

the inaccuracy caused by those problems can be simply rectified by fine-tuning.

• Both the resistance and reactance of a UHF RFID antenna are essential in deter-

mining the power transferred from the antenna into the attached chip. Hence a

good measurement of both resistance and reactance is required.

As discussed in Section 6.2, to enable maximum power transfer from a UHF RFID tag

antenna to a UHF RFID tag chip, the tag antenna should have an impedance conjugate

of the input impedance of the tag chip. Also, a tag antenna should have at least unity

gain with a wide antenna gain pattern, to enable easy absorbtion of RF energy from

any RFID reader, and transmission of that energy to the feed point.

There is no literature which explicitly discusses the measurement methods used to

determine the performance of an RFID tag antenna. The closest literature is on the

measurement of small antennas used by a mobile handset, which literature will be dis-

cussed in detail in the next section. Section 9.3 shows the experimental settings while

Section 9.4 presents the experimental results, together with some interpretation of the

results. Further modifications and improvements on measurement were carried out

and discussed in Section 9.5 while the effect of inaccuracy in measurement is presented

in Section 9.6. A conclusion is presented in Section 9.7.

This chapter is the extended version of a published paper [128].

9.2 Background and Literature Review

For a direct feed measurement using a network analyser, an RF cable is required to link

the antenna of interest to the network analyser. There is no doubt that an RF feed cable

will affect the measurement of input impedance, current distribution, and radiation

characteristic of the antenna.

Page 257

9.2 Background and Literature Review

The source of the error is the appearance of common mode current on the outside sur-

face of the outer conductor of a coaxial cable (Fig. 9.1(a)). In the case of antenna mea-

surement, when the antenna is a balanced load, this common mode current is caused

by the balanced to unbalanced connection between an antenna under test (AUT) and

the coaxial cable as shown in Fig. 9.1(b), which is the case in most of the measurements

carried out on the antenna prototypes presented in this thesis where the AUTs are of

balanced structure.

S11 will be the only antenna parameters to be measured and compared with the simu-

lation results in this chapter. Another useful antenna parameter for which comparison

between simulation and experiment may be of interest is the radiation pattern as dis-

cussed in [129]. However, we believe that S11 is more important than radiation pattern

in the case of an RFID tag antenna as the value of S11 allows the evaluation of the

power transfer from the antenna to the tag chip, and power transfer from the antenna

to the tag chip is directly proportional to the read range of an RFID tag. Although it

is good to have the verification of simulated radiation pattern, a tag antenna gain is

averagely unity in all direction and the orientation of a tag antenna with respect to an

RFID reader antenna is random in real life. This will reduce the importance of radiation

pattern of an RFID tag antenna in predicting its read range.

I

I

I

I

I’

I’

(a)

V

(b)

Figure 9.1. Explanation of common mode current. (a) Common mode current flows on the

outside surface of the outer conductor of a coaxial cable (b) Common mode current

caused by the direct connection of a balanced load to an unbalanced connecting cable.

DeMarinis [130] proposed the use of a ferrite choke as an effective solution while Icheln

[131, 132] proposed the use of Balun to transfer the unbalanced nature of a coaxial

Page 258


cable to a balanced antenna under test (AUT) to minimise the effect of the RF cable.

However, Icheln stressed that the use of Balun outperformed the use of ferrite choke

despite the fact that Balun is often band limited. The reason given was that ferrite

chokes are often lossy and hence affect the measurement of the radiation pattern, which

a properly designed Balun does not.

Alternative methods include a measurement scheme using a microstrip transmission-

line to connect a coaxial cable to an AUT as suggested by Chen [133], and also, the

use of a stack of metal rings to suppress the appearance of common mode current as

presented by Kahng [134].

9.3 Settings and Connection

An RFID tag antenna is normally made of thin copper tracks on an adhesive sheet.

This is to make an RFID tag easily and readily attached to an object. However, for

convenience, the prototype tag antennas under test are fabricated on a solid dielectric

material to have a solid foundation. The material chosen is FR4. An SMA connector

(sometimes with reduced edges for the reason explained later in this section) is then

used as the feed point of the AUT in order to connect the AUT to a network analyser

for measurement. Measurement results are then compared with the simulated results

obtained through Ansoft HFSS simulation software. In all cases, the simulation results

are used as the standard results.

Three types of coaxial cable configuration as shown in Fig. 9.2 were used for testing

purposes:

1. RG223 coaxial cable.

2. RG223 coaxial cable with ferrite sleeve.

3. Gore GMCA 190-1 (now known as G2) coaxial cable.

Two types of antenna were tested using some or all of the coaxial configurations stated

above:

Page 259

9.3 Settings and Connection

1

2

3

Figure 9.2. Cables used for measuring experiment. (1) RG223 coaxial cable. (2) RG223 coaxial

cable with ferrite sleeve. (3) Gore GMCA 190-1 coaxial cable.

1. A balanced bow tie antenna.

A balanced bow tie antenna is fabricated and is shown in Fig. 9.3. The feed point

of the antenna (with an SMA connector) is from the back and through the FR4

material as shown in Fig 9.3 (a).

(a) Picture of the bow tie

antenna

Side View(i) (ii)

Front View

Copper

FR4

SMA

ReducedEdgesVia

Holes

(b) Details and side view of the bow tie antenna

Figure 9.3. A bow tie antenna on FR4. (a) A picture showing the bow tie antenna used in

measurement experiments. (b) Front view and side view of the bow tie antenna with

(i) SMA connector (ii) modified SMA connector to enhance measurement results.

Page 260


2. Half of the balanced bow tie antenna on ground plane.

The dimensions of the balanced bow tie antenna (Fig. 9.3) are made used of in

fabricating this special antenna, which is half of the balanced bow tie antenna

on a ground plane (Fig. 9.4). This unbalanced version of the bow tie antenna is

soldered on an SMA connector, which is mounted on a ground plane. The idea is

to combine the unbalanced bow tie antenna with its image created by the ground

plane, to simulate a complete balanced bow tie antenna. It has been proven in

theory that a half unbalanced bow tie antenna on a ground plane will have half

of the input impedance of a complete balanced bow tie antenna.

(a) Picture of the half bow tie

antenna

(b) Details and side view of the half bow

tie antenna

Figure 9.4. Half bow tie antenna on ground plane. (a) A picture showing the half bow tie

antenna on ground plane used in measurement experiments. (b) Side view of the half

bow tie antenna on ground plane. The SMA connector does not need to modified

(edged) like the one shown in Fig. 9.3(b) as it is completely shielded by the metallic

ground plane.

The initial measurement results to illustrate the effect of the SMA connector of the

balanced bow tie antenna are shown in Fig. 9.5, where a comparison between the sim-

ulated and measured reactance of a simple bow tie antenna is made.

These initial measurement results show that the SMA feed arrangement of the balanced

bow tie antenna has a significant impact on the results. The size of the SMA connector

is not negligible in relation to the size of the AUT. The edges of this SMA connector

provide an alternate path for displacement current. The solution is to minimise the

edge of the SMA connector as shown in Fig 9.3 (b).

Page 261

9.4 Measurement Results

Also, from Fig. 9.5, it is shown that at a higher frequency, the accuracy of the mea-

surement of reactance drops. This is consistent with the original expectation, where an

accurate measurement is harder to obtain with the increment of frequency.

To avoid the influence of the SMA connector on the measuring results (as the main aim

is to investigate the effect of connecting cables), in this chapter, all the measurements

were carried out with a modified SMA connector, unless the AUT is the half bow tie

antenna on ground plane or otherwise stated.

Comparison of Simulated and Measured Reactance Value

-40

-20

0

20

40

60

80

100

1.1 1.25 1.4

Frequency (GHz)

Re

ac

tan

ce

(O

hm

)

Simulated

Modified SMA

SMA

Figure 9.5. The effect of SMA connector on measurement results. The edges of the SMA

connector increase the capacitance value and hence lower the overall reactance values.

The modified (edged) SMA connector improves the measurement results significantly.

Cable (1) in Fig. 9.2 was used for all these cases.


9.4.1 A Balanced Bow Tie Antenna

As mentioned before, simulation results are used as the standard results. The inaccu-

racy of measurements using Cable (1), Cable (2) and Cable (3) are on average similar

as shown in Fig. 9.6 and Fig. 9.7. However, just by touching the cables with a person’s

Page 262


hand and monitoring the impedance of the AUT, it is confirmed that Cable (1) and Ca-

ble (3), without the addition of ferrite sleeves, are more susceptible to the environment.

Resistance Comparison

14

16

18

20

22

24

26

28

30

860 870 880 890 900 910 920 930 940 950 960

Frequency (MHz)

Resis

tan

ce (

Oh

m)

Simulated

Cable (1)

Cable (2)

Cable (3)

With Ground Plane

With Balun

Figure 9.6. Simulated and measured resistance of AUT. Simulation results are used as the

standard results.

Reactance Comparison

-75

-70

-65

-60

-55

-50

-45

-40

-35

-30

-25

860 870 880 890 900 910 920 930 940 950 960

Frequency (MHz)

Reacta

nce (

Oh

m) Simulated

Cable (1)

Cable (2)

Cable (3)

With Ground Plane

With Balun

Figure 9.7. Simulated and measured reactance of AUT. Simulation results are used as the

standard results.

Page 263


9.4.2 Half Bow Tie Antenna on Ground Plane

It is observed that the ground plane which the half bow tie antenna is mounted on

provided sufficient shielding between the cable and the AUT. The results obtained

using any of the cable shows negligible difference.

In all comparisons and graphs presented in this section, measured values for a half bal-

anced bow tie antenna on ground plane have been doubled for the reason explained

in the previous section (Section 9.3) to enable easy comparison. In comparison of the

resistance values in Fig. 9.6, the measured results are constantly higher than the sim-

ulated results. This is most probably due to the loading effect of the ground plane.

Nonetheless, in comparison of the reactance values in Fig. 9.7, the measured results

are slightly better than all the results obtained on a balanced bow tie antenna, with

the exception of Cable(3), which is a high quality shielded coaxial cable. Hence, the

ground plane has effectively shielded the cable and the measuring instrument from

any unwanted coupling.

The challenge in using this method is that the AUT must be symmetrical in nature,

which is not always the case for an RFID tag antenna. Also, the distance of the half

bow tie antenna from the ground plane must be half of the distance between the two

feed points of a balanced bow tie antenna.

9.4.3 Results Interpretation

In the real part (resistance) comparison in Fig. 9.6, the half bow tie antenna on ground

plane configuration is the only measurement that gives a higher resistance as compared

to other measurement configurations when compared with the simulation results. This

is probably caused by the loading effect of the finite metal ground plane.

For the imaginary part (reactance), All measurement configurations matched well in

the region 900 to 920 MHz, where the reactance values fall within a 5 Ω range.

The position of ferrite beads affects the measurement results according to [135]. A

small test showed that ferrite beads should not be located too near to the feed point of

Page 264


the antenna. It will increase the measured resistance, as the ferrite beads will act as an

additional load.

Proper usage of a ferrite sleeve cable (Cable (2)), improves the measurement of resis-

tance quite significantly. However, there is not much improvement in the measurement

of reactance using this method.

9.5 Discussion and Improvement

Using coaxial cable with high shielding can improve the measurement results by re-

ducing the coupling between the cable and the AUT. However, not all errors can

be eliminated as inaccuracy may be introduced by the balanced to unbalanced is-

sue. Hence a Balun was designed by using coplanar strips with an RF transformer

(Fig. 9.8(a)). The model of the chosen commercial RF transformer is TC1-1-13M with

with characteristic impedance of 50 Ω and with operating frequency range of 4.5 to

3000 MHz.

The biggest challenge in using an RF transformer is to have 50 Ω coplanar strips. Ac-

cording to [86], the impedance of coplanar strips is physically limited to above 45 Ω.

Even for 50 Ω coplanar strips, the fabrication process is extremely difficult. Since the

AUT is small, the track size W1 and W2 cannot be too wide (W1 and W2 < 5 mm). To

obtain 50 Ω coplanar strips with W1 and W2 less than 5 mm, the width of separation,

S, has to be around 0.05 mm for FR4 material with εr equals 4.4. The final design uses a

more expensive composite dielectric consisting of polytetraflouroethylene (teflon) and

ceramic, with an εr of 10.2. W1 and W2 are both 2 mm in width and the separation is

0.12 mm. The calculated characteristic impedance is 49.35 Ω.

The measurement of an impedance using the network analyser requires the establish-

ment of a reference plane, and this is normally performed by placing a short-circuit at

the end of the measurement cable, with all parts of the measurement cable up to that

point being of uniform 50 Ω characteristic impedance.

However for some of the measurements undertaken, there is some non-uniformity of

structure between the end of the coaxial cable and the pair of antenna terminals at

Page 265

9.5 Discussion and Improvement

which we would like to establish a reference plane. If during calibration we place a

shorting pin across the antenna terminals will have, because of the small size of the

pin, not a complete short-circuit but some small inductance.

To remove the effect of this inductance from the measurement, we employ the nor-

malising function of the network analyser that has the effect of making this residual

inductance appear as a perfect short-circuit. Then, without further normalising of the

analyser, the short-circuit is removed and the measurement is taken.

These steps are implemented with the use of a special shorting connection, i.e. some-

thing that resembles the AUT but with a shorting pin across the positive and negative

terminals.

This method is considered to be a practical but approximate method of dealing with the

problems of not being able to maintain a uniform 50 Ω characteristic impedance in a

measurement cable up to the antenna terminals and not being able to place a complete

short-circuit across those terminals.

The results obtained using a Balun are summarised in Table 9.1. It can be seen that

with the use of Balun, the measurement of resistance is the best, though not much

better, than the rest. On the other hand, a high quality co-axial cable provides the best

results for reactance measurement.

Table 9.1. Average error and maximum error. Eave is the average error; Emax is the maximum

error.

Eave (Ω) Emax (Ω)

Re Im Re Im

Cable (1) 2.55 3.38 3.72 6.49

Cable (2) 3.29 3.64 3.96 8.03

Cable (3) 4.91 0.59 6.38 1.62

With Ground Plane 2.62 1.63 4.04 2.91

With Balun 2.35 4.12 3.90 5.26

It is inconclusive as to which measurement method is the best among all, as a direct

measurement using a normal coaxial cable offers comparably acceptable results. As

Page 266


AUT

RF Transformer

NetworkAnalyser

(a)

S

W

t

W

h

1 2

(b)

(c)

Figure 9.8. The designed Balun. The use of Balun to connect a balanced load to unbalanced

coaxial cable: (a) Top view of the design. (b) The design parameters of coplanar strips.

(c) A picture where the designed Balun is connected to a AUT.

can be seen from Fig. 9.6 and Fig. 9.7, the worst results are still within 10 Ω for the both

resistance and reactance measurements within 860 - 960 MHz band.

Note that the results are only applicable for small RFID antennas. Larger antennas,

such as a half-wave dipole, will most certainly require a Balun or ferrite sleeve for

proper measurement.

Again, in all the comparisons, simulated results are used as the standard values. It is

assumed that with proper fabrication process, an actual antenna with similar charac-

teristics with the modelled antenna, can be obtained. In short:

1. Measurement accuracy drops as frequency increases.

2. From the results shown, a ferrite sleeve does improve the measurement results.

However, it is discovered that the placement of the ferrite sleeve especially the

distance away the ferrite sleeve from the antenna under test, may affect the mea-

surement results.

Page 267

9.6 Effect of Inaccuracy in Measurement

3. Half of the AUT on ground plane gives best improvement overall. However, only

some antennas can be cut into half, such as those that are symmetrical at the feed

point.

4. A shorted antenna, rather then a normal shorted SMA connector, is used for short

circuit calibration of a network analyser before any measurement is carried out.

The investigation of RF cable effects on RFID tag antenna impedance measurement

exposes the difficulties in getting accurate measurement and obtains the best match

between AUT and chip. In actual RFID tag design, it is best to:

1. Design by simulation. Adjust the dimensions of the antenna so that the input

impedance equals to the conjugate of the chip impedance.

2. Fabricate the designed antenna. Measure the antenna input impedance.

3. Attach the tag chip on the antenna. Fine-tune the antenna, such as by reducing

the dimension of certain part of the antenna, while monitoring the read range of

the tag.


In the previous section, it was inconclusive as to which measuring method is the best.

However, with the proper experimental setup, the measurement error when compared

with the simulated results is within a small margin. In this section, the effect of mis-

match caused by the inaccuracy in measurement of the antenna impedance is pre-

sented. The maximum power transfer equation is used to examine this effect, and

the results are presented in graphical form. The analysis presented can also be applied

to any degree of mismatch, as it is not limited to just a small mismatch.

Maximum power transfer is introduced in Section 6.2. When there is no conjugate

match between the RFID antenna and the chip, the power loss can be computed through:

PlostPavailable

=∣∣∣∣Zant − Z∗

cctZant + Zcct

∣∣∣∣2

(9.1)

Page 268


where Zcct is the impedance of chip and Zant is the impedance of AUT. When Zcct

equals Zant, (9.1) will be zero, or in other words, all the available power is transferred

from the antenna to the chip and there is no power loss.

Both the impedances, Zcct and Zant can be expressed in terms of resistance and reac-

tance:

Zcct = a + jb (9.2)

Zant = c + jd (9.3)

If the reactance of the AUT and the chip are equal and opposite (b = −d), (9.1) can be

simplified to:

PlostPavailable

=∣∣∣∣ a − ca + c

∣∣∣∣2

(9.4)

If resistance of the AUT and the chip are equal (a = c), (9.1) becomes:

PlostPavailable

=∣∣∣∣ (b + d)2 + j2a(b + d)

4a2 + (b + d)2

∣∣∣∣2

(9.5)

or in magnitude form:

PlostPavailable

=(b + d)4 + 4a2(b + d)2

[4a2 + (b + d)2]2(9.6)

Note that b is always negative as an RFID tag chip is always capacitive in nature due to

the reservoir capacitor located in the rectifier circuit. Hence the term (b + d) is actually

the difference between the magnitudes of Im(Zcct) and Im(Zant). This term should be

minimised for maximum power transfer as maximum power transfer occurs when b

equals −d (when (b + d) equals zero), or in other words Zcct is the conjugate of Zant.

The power transfer efficiency can be computed through:

Page 269


η = 1 − PlostPavailable

(9.7)

Power Transfer Efficiency (Equal Reactance)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

-40 -30 -20 -10 0 10 20

Difference in Resistance (Rcct-Rant) (Ohm)

Eff

icie

ncy

20

15

10

5

Figure 9.9. The power transfer efficiency between AUT and chip when reactances are equal

and opposite. This graph is based on (9.4) and (9.7). The chip input resistance is

indicated by the small number besides each traces in the graph.

The plots of (9.7) with PlostPavailable

given as (9.4) and (9.6) are presented as Fig. 9.9 and

Fig. 9.10 respectively. These graphs serve two purposes. Firstly, they show the power

transfer loss when there is a mismatch in either the real part or imaginary part between

the AUT and the chip input impedance. It can also be seen from both graphs (Fig. 9.9

and Fig. 9.10) that the resistance of a UHF chip has great impact on the power trans-

fer efficiency. A UHF chip with a higher resistance value has a better power transfer

efficiency when compared with a UHF chip with a lower resistance value, though the

degree of mismatch is the same for these two cases.

The second purpose is to show the impact when the characteristics of the modelled

antenna are not accurately translated into the fabricated antenna. If the initial antenna

design is entirely carried out using a simulation program, this problem will most likely

happen. It shows that by using a higher resistance value chip, the power transfer loss

can be minimised.

Page 270


Power Transfer Efficiency (Equal Resistance)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

-30 -20 -10 0 10 20 30

Reactance Difference (Xcct-Xant) (Ohm)

Eff

icie

ncy

20

15

10

5

2

Figure 9.10. The power transfer efficiency between AUT and chip when resistances are

equal. This graph is based on (9.6) and (9.7). The chip input resistance is indicated

by the small number besides each traces in the graph.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

-30 -20 -10 0 10 20

Difference in Both Resistance and Reactance (Ohm)

Eff

icie

ncy

Estimation

Power Loss Curve

Figure 9.11. The comparison between the estimated and calculated values on the power loss

curve. The estimated values are obtained by multiplying (9.4) and (9.6). Calculated

values are based on (9.1). The resistance of the chip is set to 20 Ω in this case.

Page 271

9.7 Conclusion

In the case where there are mismatches in both resistance and reactance (which is in

fact a common scenario), the power transfer efficiency can be computed using (9.1).

Alternatively it can be approximated, even if not very accurately, by multiplying the

two power transfer loss equations (9.4) and (9.6).

An example is shown as Fig. 9.11. In this example, the resistance of the chip (Rcct) is

chosen to be 20 Ω. The x-axis shows the difference of both resistance (Rcct - Rant) and

reactance (Xcct - Xant). For example, at a point in the graph corresponding to value -10

of the x-axis, it simply means that the difference the resistance and reactance between

a tag chip and a tag antenna are both 10 Ω. Note that this graph does not correspond

to the multiplication of Fig. 9.9 and Fig. 9.10, as those graphs either assumed equal

resistance or equal and opposite reactance.

9.7 Conclusion

Detailed investigation of RF cable effects on UHF RFID tag antenna impedance mea-

surement was presented in this chapter. It offers an insight on the difficulties in ob-

taining the correct measurement of impedance. Measurements were carried out using

several different configurations. It is shown that it is impossible to obtain the exact sim-

ulation results (the ideal case). Nonetheless, the impact of the inaccuracy of impedance

measurement on the maximum power transfer between an antenna and a chip can be

predicted. As long as a tag antenna is designed to be able to fine-tuned easily, the chal-

lenges in measuring an antenna’s impedance should not impede the development and

deployment of an RFID tag antenna.

Page 272

Chapter 10

Thesis Conclusions

CONCLUSIONS of this thesis are presented in this chapter.

It reviews and summarises all the contributions from the re-

search during the candidature period. Also, some possible

future research extensions are discussed. It is hoped that this research will

contribute to the development of RFID technology until the vision of ubiq-

uitous identification is achieved.

Page 273

10.1 Introduction

10.1 Introduction

This chapter presents the conclusions of this thesis. A brief overall review is compiled

as Section 10.2. A few interesting research extensions are discussed in Section 10.3.

The summary of the contributions that arise from the research work published by this

thesis, are listed in Section 10.4, followed by the conclusion of this thesis.

10.2 Review of Thesis

The title of this thesis is “Antenna Positioning Analysis and Dual-Frequency Antenna

Design of High Frequency Ratio for Advanced Electronic Code Responding Labels”.

As explained in the introduction chapter of this thesis (Chapter 1), this thesis consists

of two major parts:

1. Antenna positioning analysis.

2. RFID tag antenna design (including novel dual-frequency antenna).

For the first part (Chapter 3 to Chapter 5), research work was devoted to the study

of path loss models and the development of a simulation software to investigate the

signal strength of RFID readers within a certain area. The developed software was

varied to contribute to numerous areas (Refer Section 10.4).

It is regrettable that actual reader deployment in a warehouse was not able to be carried

out during the research period due to a lack of RFID readers. Nonetheless small-scale

testing (one to two readers) has shown that the model suggested gives good estimation

of reader antenna signal strength.

In the second part (Chapter 6 to Chapter 9), designs of HF and UHF RFID antennas

for various new industrially driven applications are described. This is followed by

the design and testing of novel dual-frequency RFID tag antenna, which operates at

both HF and UHF. Also included are the measuring techniques used when measuring

antenna impedances during the prototyping stage.

Page 274

Chapter 10 Thesis Conclusions

The research work on tag antenna designs covers the theoretical tag antenna design,

tag antenna simulation, tag antenna prototyping and measurement, the manufacturing

of fully functional RFID tags and on-site testing. Optimisation of antenna size and

read range performance was carried out when possible, but is not the priority of the

research, as the main aim in those sections was to show feasibility and flexibility in the

designing of tag antennas for hostile environments.

Also it is unavoidable that when comparing free space read range (not in hostile en-

vironment), the RFID tags with novel antennas designed and presented in this thesis

are outperformed by commercial tags. Nonetheless, normal commercial tags would

not be operational in a hostile environment while the newly designed tags can offer an

acceptable read range performance.

Overall, this thesis provides an in-depth investigation of collision problems in dense

reader environments, the tag attachment problem in hostile environments, and the

problem of development of a compact dual-frequency antenna of high frequency ratio.

This work hopefully will contribute to the engineering community which is working

towards an ubiquitous identification system in supply chains.

10.3 Possible Research Extensions

The contents presented in this thesis covers the work carried out within the three-year

candidature time span. Though all the research aims have been achieved with results

and findings published in international level conferences and journals, there are still

some interesting topics for further study. Not really related to the studies of antenna

design and electromagnetic, and in no particular order of importance they are:

1. Real-time Simulator

The core of the simulation software, which compute the signal strength of reader

antennas has been developed and presented in this thesis. However, the develop-

ment of a real-time simulator can be beneficial to real-time adjustment of antenna

placement in a dense RFID environment.

Page 275

10.4 Summary of Contributions

2. Dual-frequency RFID Protocol

The proposed dual-frequency RFID tag antenna will work in HF (using ISO/IEC

18000-3:2004 [33]) and UHF (EPCglobal C1G2 [35]). This research extension is to

examine both the above stated protocols and to develop a combined protocol.

3. Dual-frequency Chip Design

Prototypes of dual-frequency antennas have been proven functioning at both the

HF and UHF bands. However, there is no available dual-frequency chip operat-

ing at these two bands. The design of a dual-frequency chip will require exper-

tise in semiconductor design and RFID design, in integrating the HF standards

(ISO/IEC 18000-3:2004) [33], and the UHF standards (EPCglobal C1G2) [35] to

produce a single chip with a single feed point.

4. RFID Cost Reduction

This a research area in manufacturing engineering, in which new manufacturing

techniques are to be developed based on creative choice of materials.

10.4 Summary of Contributions

1. Antenna Positioning Analysis

A comprehensive study of the RFID collision problems (which focus on the Reader

Interference with Tag Replies (RITR) problem and the Regulatory Reader Shut-

down (RRS) problem), and the development of a path loss model to predict signal

strength radiated by RFID reader antennas was carried out. A piece-wise linear

path loss model is suggested for simulation purposes.

2. Development of Simulation Software

Software written in MATLAB is presented. This software simulates the signal

strength of RFID reader antennas in a defined zone. The simulated results can

be used for reader antenna placement to minimise interference between reader

antennas, maximise coverage, while adhering to local regulations.

Page 276

Chapter 10 Thesis Conclusions

3. Investigation of Sources of Simulation Errors

Intensive research was carried out to identify the common sources of simula-

tion errors, especially in the case of RFID reader antenna simulation in a defined

vicinity. Guidelines were presented to help anyone who is interested in develop-

ing simulation software for RFID deployment.

4. Reader Synchronisation

The idea of reader synchronisation is examined. Reader synchronisation is in-

troduced to allow better RFID performance and coverage when strict regulations

are imposed on RFID deployment in European countries. Various methods were

discussed with several new fine-tuning techniques presented.

5. Regulation and Human Safety

Work was carried out to simulate the radiation power density in the surround-

ings of a dense RFID reader environment. This research was primarily a consul-

tancy on Specified Absorption Rate (SAR). The report generated was essential to

support the experimental work carried out for the approval of 4 W EIRP Radio

Frequency Identification (RFID) license in Australia.

6. Threshold of Second Carrier Sensing

This is an extension study of an existing newly proposed protocol to reduce the

Reader Induced Tag Confusion (RITC) problem in a dense reader environment.

Based on simulation results, recommendations were made on the threshold val-

ues for second carrier sensing with various antenna configurations and place-

ments.

7. HF and UHF RFID Tag and System Design

Pallet level HF and UHF RFID tag antennas are readily available. However, there

are still huge challenges in item level tagging for many types of consumer items,

such as items with ionised liquid (water) or with metallic parts. Throughout

the research candidature, several case studies were completed with the designs,

prototyping and testing of industrially driven RFID HF and UHF tag antennas.

Page 277

10.5 Conclusion

8. Investigation on Measuring Techniques

Several measuring techniques were designed and tested to obtain the impedance

of prototype HF and UHF RFID antennas. This is a well-known problem to RFID

tag antenna designer to measure the impedance of a prototype antenna.

9. Deployment of RFID Tags in Piggery

HF and UHF RFID tags were deployed in a piggery for the first time in Australia.

This field trial demonstrates that the use of RFID technology is beneficial to the

pig industry, as when the eating behaviour of pigs is monitored, the health of

pigs can be inferred from the eating frequencies.

10. Novel Dual-Frequency RFID Tag Antenna

An RFID tag antenna which operates in both HF and UHF RFID bands was de-

signed. This first of its kind antenna has a single feed point and is ready for

wide-scale adoption when a dual-frequency RFID chip is commercially avail-

able. Being a dual-frequency antenna, it embraces both the benefits offered by

HF and UHF antennas. For completeness, several variants of this dual-frequency

antenna were explored, including one with a multi-feed point.

10.5 Conclusion

This chapter identifies all the research work presented in this thesis, which contributes

to RFID practice in recommending various reader antenna positioning and deploy-

ment techniques, design and analysis of HF and UHF tag antennas for hostile environ-

ments, and development of a novel dual-frequency RFID tag antenna.

Page 278

Appendix A

Calculation of Impedanceof Coplanar Strip Line

The characteristic impedance of a coplanar strips (CPS) is computed with reference

to [136].

The variables defined in Fig. 9.8(b) are S, W1, W2, t, h.

To simplify the mathematical expression in calculating the characteristic impedance,

two new variables are defined:

a =S2

(A.1)

b =W1 + W2 + S

2(A.2)

Characteristic impedance, Z0:

Z0 =120π√

εe f f

K(k′)K(k)

(A.3)

where:

εe f f = 1 +12(εr − 1)

K(k)K(k′)

K(k′h)K(kh)

(A.4)

k =

√1 −

( ab

)2(A.5)

k′ =√

1 − k2 (A.6)

Page 279

kh =

√1 − sinh2(πa

2h )sinh2(πb

2h )(A.7)

k′h =√

1 − k2h (A.8)

where K is the complete elliptical integral of the first kind.

Page 280

Appendix B

Some Notes on theConcept of Inductance

This appendix records in-sight obtained from discussion with Prof. Peter Cole on the

concept of inductance. It complements the material of Section 6.10.2.

1. Objectives

The objective of these notes is to clarify some concepts concerned with mutual

and self inductance. This appendix aims to give proper meaning to statements

about inductance that are commonly made without adequate clarification of what

they mean.

2. Some Basics

Inductance, either self or mutual, is defined as flux linkage per unit current. This

definition implies two closed contours: contour C1 for the path of the current,

and contour C2 to define the boundary of the flux linked. We digress to provide

a reminder that the Biot-Savart law commonly quoted for the field of a current

element seems to imply that a current element can act alone to create a field.

However, that is not so. The validity of that law is limited to the case when it

is used to calculate a field by integrating over the full path of the current, which

will flow in a closed contour which we have chosen above to label as C1. But in

this analysis we will not have to use the Biot-Savart law. We will use instead an

expression for magnetic vector potential from which the magnetic flux density

may be derived. But we believe that expression for magnetic vector potential has

its own validity restricted to the case when it is integrated over the full closed

contour of the causing current.

Page 281

3. Calculation of flux linked

The flux Φ2 linked by a contour C2 enclosing a surface S2 through which flows a

flux density B2 derivable from a vector potential A2 is given by:

Φ2 =∫

S2

B2 · ds2 =∫

S2

(curlA2) · ds2 =∮

C2

A2 · dr2 (B.1)

where we have used Stokes’ law to convert from a surface integral to a contour

integral.

Now we can calculate the magnetic vector potential A2(r2) caused by the current

flowing in contour C1 by

A2(r2) =µ0

4π

∮C1

I(r1)dr1

r12(B.2)

where r12 = |r2 − r1| is the scalar distance from the current element I1(r1) to the

potential point r2. Combining the above results and dividing by the current I1

we have the expression for the mutual inductance between the circuit of contour

C1 and the circuit of contour C2. This is

L21 =µ0

4π

∮C2

[∮C1

dr1

r12·]

dr2 (B.3)

Since the limits of the two integrals are independent, we may write this as:

L21 =µ0

4π

∮C2

∮C1

dr1 · dr2

r12(B.4)

We see that L12 = L21, and we call of this result the mutual inductance M. The

formula is known as Neumann’s formula for inductance.

4. Some mathematical difficulties

For contours that do not intersect, there are none. But when contours intersect we

have a difficulty at points where r12 = 0. Furthermore we can not in this simple

form use the result to calculate the self inductance of a circuit.

Page 282

Appendix B Some Notes on the Concept of Inductance

5. Resolution

To deal with these problems, we will reduce the level of idealisation of the prob-

lem by allowing for the non-zero dimensions of that wire carrying the current.

We illustrate in Fig. B.1 the case of a round wire of radius a carrying the current I1,

and consider the radio frequency case when the current will flow on the surface

of the wire. In the diagram we express position vectors r1 and r2 in relation to an

arbitrarily positioned origin.

r2

2a

B r( )2

Arbitrary originO

r1

Figure B.1. Illustration to explain the calculation of flux density of a round wire of radius a.

In calculating at points r2 the flux density B2(r2) and magnetic vector potential

A2(r2) we will make the approximation that the current flows not on the surface

of the wire but at the centre defined in the diagram by the position vector r1. In

calculating the flux linked, we will not let position vector r2 go inside the wire,

but will restrict it to be no nearer the centre than the radius a. Contour C1 now

follows the centre of the wire, and contour C2 now follows the inner edge. Thus

the scalar difference r12 = |r2 − r1| will be no smaller than a, and the integral can

be performed. There are formulae in the literature, for the “self inductance” of

round wires and of wires of various shapes, that are based on this principle.

6. What we find in the literature

Page 283

Among what we can find in the literature are expressions for the self inductance

per unit length of a straight wire, of the given cross sectional shape, or straight

round wire of the given radius.

What this really means is that it (the inductance per unit length) is the contribu-

tion to the integral:

L21 =µ0

4π

∮C2

[∮C1

dr1

r12·]

dr2 (B.5)

which is obtained by integrating (twice) over a length LW of straight wire, mod-

ifying the contours C1 and C2 as described above to ensure that r12 is not zero,

and dividing by the length LW . We find that as LW becomes large, the integral

over the LW part of C1 becomes independent of LW , whereas the integral over

part of C2 is proportional to LW , so a length independent result is obtained after

dividing by LW . But any statement about the “inductance” of the wire segment

must always be understood as being only a contribution to the true inductance

(self or mutual) of a complete circuit (or circuits) and that the complete induc-

tance will consist of the sum of the so-called “inductances” of all wire segments

and so-called “mutual inductances” between all segments.

Page 284

Appendix C

MATLAB Code for PathLoss Calculation

This appendix includes all the MATLAB code required to produce the simulation re-

sults shown in Chapter 3. Detailed explanation on how the code works is also included

in Chapter 3.

Section C.1 shows the main body of the simulation code, where the settings for any

output plot is located. Also, the area of simulation can be changed there. Theoretically,

there is no limit to the size of the area, except the limitation of the computer used for

simulation.

Section C.2 has a data grid to represent the antenna gain pattern for the reader antennas

used in the simulation. If a different antenna is used, a new data grid is required.

The data grid is obtained through actual antenna gain measurement in an anechoic

chamber.

Section C.3 computes the path loss based on path loss model discussed in Chapter 3.

C.1 Main Code

1 c l e a r ;

warning o f f MATLAB: divideByZero ;

3

antenna1 = antgainv2 ( 5 0 0 , 4 5 0 , 5 0 ) ;

5 antenna2 = antgainv2 ( 5 0 0 , 4 0 0 , 5 0 ) ;

t o t a l = 10∗ log10 ( 1 0 . ˆ ( 0 . 1 ∗ ( antenna1 ) ) + 1 0 . ˆ ( 0 . 1 ∗ ( antenna2 ) ) ) ;

7

Page 285

C.2 Antenna Gain Pattern

f igure ( 1 ) =pcolor ( t o t a l ) ;

9 hold ;

11 co lor1 = [0 0 0 . 5 ] ; c o lor2 = [0 0 1 ] ;

co lor3 = [0 1 1 ] ; c o lor4 = [1 1 0 ] ;

13 c o l o r 5 = [1 0 0 ] ; c o l o r 6 = [ 0 . 5 0 0 ] ;

15 colormap ( usercolormap ( color1 , color2 , color3 , color4 , color5 , c o l or 6 ) ) ;

17 c a x i s ([−80 0 ] ) ;

colorbar ;

19 s e t ( gca , ’ PlotBoxAspectRatioMode ’ , ’ manual ’ ) ;

s e t ( f igure ( 1 ) , ’ EdgeColor ’ , ’ none ’ ) ;

21 c l a b e l ( contour ( t o t a l , 2 0 , ’ k ’ ) , ’ manual ’ ) ;

hold ;

C.2 Antenna Gain Pattern

function x=antgainv2 ( size , x , y ) disp ( ’ Loading antenna gain pat te rn &

2 path l o s s . . . . ’ ) ;

n=s ize ; %a r r a y s i z e

4

a=y ; %antenna l o c a t i o n − row

6 b=x ; %antenna l o c a t i o n − column

%antenna f a c i n g down

8

gain=ones ( n , n ) ;

10

%d a t a g r i d

12

ddata=[−14 −13.8500 −13.7000 −13.5500 −13.4000 −13.2500

14 −13.1000 −12.9500 −12.8000 −12.6500 −12.5 −12.3500 −12.2000

−12.0500 −11.9000 −11.7500 −11.6000 −11.4500 −11.3000 −11.1500

16 −11 −10.8000 −10.6000 −10.4000 −10.2000 −10.0000 −9.8000

−9.6000 −9.4000 −9.2000 −9 −8.8500 −8.7000 −8.5500

18 −8.4000 −8.2500 −8.1000 −7.9500 −7.8000 −7.6500 −7.5

−7.3500 −7.2000 −7.0500 −6.9000 −6.7500 −6.6000 −6.4500

20 −6.3000 −6.1500 −6 −5.8500 −5.7000 −5.5500 −5.4000

Page 286

Appendix C MATLAB Code for Path Loss Calculation

−5.2500 −5.1000 −4.9500 −4.8000 −4.6500 −4.5 −4.3500

22 −4.2000 −4.0500 −3.9000 −3.7500 −3.6000 −3.4500 −3.3000

−3.1500 −3 −2.9000 −2.8000 −2.7000 −2.6000 −2.5000

24 −2.4000 −2.3000 −2.2000 −2.1000 −2 −1.8500 −1.7000

−1.5500 −1.4000 −1.2500 −1.1000 −0.9500 −0.8000 −0.6500

26 −0.5 −0.3500 −0.2000 −0.0500 0 .1000 0 .2500 0 .4000

0 .5500 0 .7000 0 .8500 1 1 . 1 1 . 2 1 . 3 1 . 4 1 . 5 1 . 6 1 . 7 1 . 8 1 . 9

28 2 2 . 1 2 . 2 2 . 3 2 .400 2 .500 2 .6000 2 .700 2 .8000

2 .900 3 3 .100 3 .2000 3 .300 3 .4000 3 .500 3 .6000

30 3 .700 3 .8000 3 .900 4 4 .0500 4 .1000 4 .1500 4 .2000

4 .2500 4 .3000 4 .3500 4 .4000 4 .4500 4 . 5 4 .5500

32 4 .6000 4 .6500 4 .7000 4 .7500 4 .8000 4 .85000 4 . 9 0

4 . 9 5 5 5 .0500 5 .1000 5 .1500 5 .2000 5 .2500 5 .3000

34 5 .3500 5 .4000 5 .4500 5 . 5 5 .5500 5 .6000 5 .6500

5 .7000 5 .7500 5 .8000 5 .8500 5 .9000 5 .9500 6 6 6 6

36 6 6 6 6 6 6 6 ] ;

38 ddata back =[−14 −14.1000 −14.2000 −14.3000 −14.4000 −14.5000

−14.6000 −14.7000 −14.8000 −14.9000 −15 −15.1000 −15.2000

40 −15.3000 −15.4000 −15.5000 −15.6000 −15.7000 −15.8000 −15.9000

−16 −16.2000 −16.4000 −16.6000 −16.8000 −17.0000 −17.2000

42 −17.4000 −17.6000 −17.8000 −18 −18.2000 −18.4000 −18.6000

−18.8000 −19.0000 −19.2000 −19.4000 −19.6000 −19.8000 −20

44 −20.0500 −20.1000 −20.1500 −20.2000 −20.2500 −20.3000 −20.3500

−20.4000 −20.4500 −20.5 −20.6500 −20.8000 −20.9500 −21.1000

46 −21.2500 −21.4000 −21.5500 −21.7000 −21.8500 −22 −22.2000

−22.4000 −22.6000 −22.8000 −23.0000 −23.2000 −23.4000 −23.6000

48 −23.8000 −24 −23.9000 −23.8000 −23.7000 −23.6000 −23.5000

−23.4000 −23.3000 −23.2000 −23.1000 −23 −22.9500 −22.9000

50 −22.8500 −22.8000 −22.7500 −22.7000 −22.6500 −22.6000 −22.5500

−22.5 −22.4500 −22.4000 −22.3500 −22.3000 −22.2500 −22.2000

52 −22.1500 −22.1000 −22.0500 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22

−22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22

54 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22

−22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22

56 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22 −22

−22 −22 −22];

Page 287

C.3 Path Loss

58

%c o m p u t a t i o n

60 for i =1 :n

for j =1 :n

62 r a t i o = abs ( ( j−b ) /( i−a ) ) ;

r a t i o = 2∗(90− atan ( r a t i o ) ∗180/ pi ) ;

64 r a t i o = round ( r a t i o ) ;

i f i−a ˜=0

66 i f a<=i

gain ( i , j ) = ddata ( 1 , r a t i o +1) ;

68 else

gain ( i , j ) = ddata back ( 1 , r a t i o +1) ; ;

70 end

else

72 gain ( i , j ) = ddata ( 1 , 1 ) ;

end

74 i f j−b==0&&a<=i

gain ( i , j ) = 6 ;

76 end

end

78 end

x=gain−path loss ( n , b , a , 0 . 0 1 ) ; %r e t u r n ga in

80 disp ( ’Done ’ ) ;

C.3 Path Loss

1 function x=path loss ( size , x , y , u n i t s )

3 n=s ize ; %a r r a y s i z e

ploss = ones ( n ) ;

5 d i s t a n c e = ones ( n ) ;

a=y ; %antenna l o c a t i o n − row

7 b=x ; %antenna l o c a t i o n − column

9 %G e n e r a t e D i s t a n c e

for i = 1 : n

11 for j = 1 : n

d i s t a n c e ( i , j ) =( sqr t ( ( a−i ) . ˆ 2 + ( j−b ) . ˆ 2 ) ) ;

Page 288

Appendix C MATLAB Code for Path Loss Calculation

13 d i s t a n c e ( i , j ) = d i s t a n c e ( i , j ) .∗ u n i t s ;

end

15 end

17 for i = 1 : n

for j = 1 : n

19

i f d i s t a n c e ( i , j ) ==0

21 d i s t a n c e ( i , j ) =1.∗ u n i t s ;

end

23

25 i f d i s t a n c e ( i , j )<8

ploss ( i , j ) =32+25∗ log10 ( d i s t a n c e ( i , j ) ) ;

27 else

ploss ( i , j ) =23+35∗ log10 ( d i s t a n c e ( i , j ) ) ;

29 end

end

31 end

33 x = ploss ;

Page 289

Appendix D

MATLAB Code for HFSSVB Script Generation

This appendix shows the MATLAB code written to generate HF coils of different size

to be simulated in Ansoft HFSS. The code based on the API produced by [107].

1 function makecoil ( turn , L , L1 , th ick , height ,w, g , s a v e l o c )

3 % add p a t h s t o t h e r e q u i r e d m− f i l e s .

addpath C: \ Matlab\ h f s s a p i \3dmodeler ;

5 addpath C: \ Matlab\ h f s s a p i \general ;

addpath C: \ Matlab\ h f s s a p i \ a n a l y s i s ;

7 addpath C: \ Matlab\ h f s s a p i \boundary ;

9 % Temporary F i l e s . Thes e f i l e s can be d e l e t e d a f t e r t h e o p t i m i z a t i o n i s

c o m p l e t e . We have t o s p e c i f y t h e c o m p l e t e pa th f o r a l l o f them .

tmpPr jF i le = ’C:\ Matlab\ h f s s a p i \temp . h f s s ’ ;

11 tmpDataFile = s t r c a t ( ’C:\ Matlab\ h f s s a p i \ ’ , s a v e l o c ) ;

t m p S c r i p t F i l e = ’C: \ Matlab\ h f s s a p i \temp . vbs ’ ;

13

% HFSS E x e c u t a b l e Path .

15 hfssExePath = ’C: \ Progra ˜1\ Ansoft\HFSS9\ h f s s . exe ’ ;

17 % C r e a t e a new temporary HFSS s c r i p t f i l e .

f i d = fopen ( tmpScr iptF i le , ’wt ’ ) ;

19

% C r e a t e a new HFSS P r o j e c t and i n s e r t a new d e s i g n .

21 hfssNewProject ( f i d ) ;

h f s s I n s e r t D e s i g n ( f id , ’ t e s t b o x ’ ) ;

23

Page 291

% C r e a t e t h e C o i l .

25

% v a r i a b l e s :

27 n = turn ;

length0 = L ;

29 length1 = L1 ;

t = t h i c k ;

31 h = height ;

width = w;

33 gap = g ;

35 % c r e a t e s u b s t r a t e

sub x = 2∗ length1 +2∗(n+2) ∗ ( gap+width ) ;

37 sub y = 2∗ length0 +2∗(n+2) ∗ ( gap+width ) ;

39 hfssBox ( f id , ’ Sub ’ ,[− sub x/2,−sub y / 2 , 0 ] , [ sub x , sub y ,−h ] , ’mm’ , ’ FR4 epoxy ’ , ’

t rue ’ ) ;

% f u n c t i o n h f s s C y l i n d e r ( f i d , Name , Axis , Center , Radius , Height , Units )

41 h fssCyl inder ( f id , ’ Via 1 ’ , ’Z ’ , [ width /2 , length0+width /2 , 0 ] , width /3 , −h ,

’mm’ , ’vacuum ’ , ’ t rue ’ ) ;

hfssCyl inder ( f id , ’ Via 2 ’ , ’Z ’ , [ width /2 , length0+width/2+n∗ ( width+gap ) ,

0 ] , width /3 , −h , ’mm’ , ’vacuum ’ , ’ t rue ’ ) ;

43 h f s s S u b t r a c t ( f id , ’ Sub ’ , ’ Via 1 ’ , ’ Via 2 ’ ) ;

45 % c r e a t e c o i l

Count=n∗4+1;

47

% Setup

49 S t a r t L o c =[0 , length0 , 0 ] ; Curr Loc= S t a r t L o c ; Ct =1;

51 while ( t rue ) Arm = s t r c a t ( ’Arm ’ , i n t 2 s t r ( Ct ) ) ;

53 i f Ct==1

%1 s t Arm

55 temp length=length1 ;

temp width=width ;

57 hfssBox ( f id ,Arm, Curr Loc , [ length1 , width , t ] , ’mm’ , ’ copper ’ , ’ f a l s e ’ ) ;

Page 292

Appendix D MATLAB Code for HFSS VB Script Generation

e l s e i f Ct>=Count

59 %f i n a l arm

temp length =(n+1)∗width+length1 +(n−1)∗gap ;

61 temp width=width ;

hfssBox ( f id ,Arm, Curr Loc , [ temp length , temp width , t ] , ’mm’ , ’ copper ’ , ’

f a l s e ’ ) ;

63 break ; else

%t o p arm

65 temp length = ( ( Ct−1)/2)∗width+2∗ length1 +( Ct−3)/2∗gap ;

temp width=width ;

67 hfssBox ( f id ,Arm, Curr Loc , [ temp length , temp width , t ] , ’mm’ , ’ copper ’ , ’ f a l s e ’ ) ;

end

69

Ct=Ct +1; Curr Loc=Curr Loc +[ temp length , temp width , 0 ] ;

71 Arm = s t r c a t ( ’Arm ’ , i n t 2 s t r ( Ct ) ) ;

73 %r i g h t arm

temp length=width ; temp width =( Ct /2)∗width+2∗ length0 +( Ct/2−1)∗gap ;

75 hfssBox ( f id ,Arm, Curr Loc , [ temp length ,− temp width , t ] , ’mm’ , ’ copper ’ , ’ f a l s e ’ )

;

77 Ct=Ct +1;

Curr Loc=Curr Loc +[ temp length ,− temp width , 0 ] ;


81 %down arm

temp length = ( ( Ct−1)/2)∗width+2∗ length1 +( Ct−3)/2∗gap ;

83 temp width=width ;

hfssBox ( f id ,Arm, Curr Loc ,[− temp length ,− temp width , t ] , ’mm’ , ’ copper ’ , ’ f a l s e ’

) ;

85

Ct=Ct +1; Curr Loc=Curr Loc+[− temp length ,− temp width , 0 ] ;


89 %l e f t arm

temp length=width ; temp width =( Ct /2)∗width+2∗ length0 +( Ct/2−1)∗gap ;

Page 293

91 hfssBox ( f id ,Arm, Curr Loc ,[− temp length , temp width , t ] , ’mm’ , ’ copper ’ , ’ f a l s e ’ )

;

Ct=Ct +1;

93 Curr Loc=Curr Loc+[− temp length , temp width , 0 ] ;

end

95

% f u n c t i o n h f s s C y l i n d e r ( f i d , Name , Axis , Center , Radius , Height , Units )

97 h fssCyl inder ( f id , ’ Via Cu 1 ’ , ’Z ’ , [ width /2 , length0+width /2 , 0 ] , width /3 , −h , ’mm’ , ’ copper ’ , ’ f a l s e ’ ) ;

h fssCyl inder ( f id , ’ Via Cu 2 ’ , ’Z ’ , [ width /2 , length0+width/2+n∗ ( width+gap ) ,

0 ] , width /3 , −h , ’mm’ , ’ copper ’ , ’ f a l s e ’ ) ;

99

%Underpass

101 hfssBox ( f id , ’ Underpass 1 ’ , [ 0 , length0 ,−h ] , [ width , width ,− t ] , ’mm’ , ’ copper ’ , ’

f a l s e ’ ) ;

hfssBox ( f id , ’ Underpass 2 ’ , [ 0 , length0+gap+width ,−h ] , [ width , n∗width +(n−1)∗gap

,− t ] , ’mm’ , ’ copper ’ , ’ f a l s e ’ ) ;

103

%Merging Cu

105 temp str= ’ ’ ;

107 for i = 1 : ( n∗4+1)

temp str= s t r c a t ( temp str , ’Arm ’ , i n t 2 s t r ( i ) ) ;

109 i f i ˜= ( n∗4+1)

temp str= s t r c a t ( temp str , ’ , ’ ) ;

111 end end

113 h fssUni te ( f id , ’ Underpass 1 ’ , ’ Underpass 2 ’ , ’ Via Cu 1 ’ , ’ Via Cu 2 ’ ,

temp str ) ;

115 %Lump p o r t

hfssRec tangle ( f id , ’ Rec t 1 ’ , ’Z ’ , [ 0 , length0+width ,−h ] , width , gap , ’mm’ ) ;

117

hfssAssignLumpedPort ( f id , ’ LumpedPort ’ , ’ Rect 1 ’ , [ width /2 , length0+width ,−h

] , [ width /2 , length0+width+gap,−h ] , ’mm’ , 5 0 , 0 ) ;

119

% Add an AirBox .

Page 294

Appendix D MATLAB Code for HFSS VB Script Generation

121 hfssBox ( f id , ’ AirBox ’ , [−1.1∗ sub y , −1.1∗ sub y , −1.1∗ sub y ] , [ 2 ∗ 1 . 1 ∗ sub y

, 2 ∗1 . 1∗ sub y , 2∗1 .1∗ sub y ] , ’mm’ , ’vacuum ’ , ’ t rue ’ ) ;

123 % Add PML Layer

%f p r i n t f ( f i d , ’ oModule . CreatePML Array (” UserDrawnGroup : = ” , f a l s e , \n ’ ) ;

125 %f p r i n t f ( f i d , ’” PMLFaces : = ” , Array ( 1 2 9 9 , 1 3 0 0 , 1 3 0 1 , 1 3 0 2 , 1 3 0 3 , 1 3 0 4 ) , \n ’ ) ;

%f p r i n t f ( f i d , ’” C r e a t e C o r n e r O b j s : = ” , t rue , \n ’ ) ;

127 %f p r i n t f ( f i d , ’” T h i c k n e s s : = ” , ”80mm” , ” RadDist : = ” , ”20mm” , \n ’ ) ;

%f p r i n t f ( f i d , ’” UseFreq : = ” , t rue , ” MinFreq : = ” , ”0 .01356GHz”) \n ’ ) ;

129

% Add a S o l u t i o n Setup .

131 %h f s s I n s e r t S o l u t i o n ( f i d , ’ Setup13 56MHz ’ , 13560000/1 e9 ) ;

%h f s s I n t e r p o l a t i n g S w e e p ( f i d , ’ Sweep10to20MHz ’ , ’ Setup13 56MHz ’ , 10000000/1

e9 , 20000000/1 e9 , 201) ;

133

135 % Save t h e p r o j e c t t o a t emporary f i l e and s o l v e i t .

%h f s s S a v e P r o j e c t ( f i d , t m p P r j F i l e , t r u e ) ;

137 %h f s s S o l v e S e t u p ( f i d , ’ Setup13 56MHz ’ ) ;

139 % Expor t t h e d a t a

%hf s s E x p o r t N e t w o r k D a t a ( f i d , tmpDataFi l e , ’ Setup13 56MHz ’ , ’ Sweep10to20MHz ’ ) ;

141

% C l o s e t h e HFSS S c r i p t F i l e .

143 f c l o s e ( f i d ) ;

145 disp ( ’VBS f i l e ready . ’ ) ;

%h f s s E x e c u t e S c r i p t ( h f s s E x e P a t h , t m p S c r i p t F i l e ) ;

Page 295

Appendix E

MATLAB Code forInductance Calculation

This appendix includes all the MATLAB code required to compute the inductance of

planar metal track.

Section E.1 is the core of the program. It will accept parameters which will define the

dimensions of a square loop antenna, including number of turns. It will then con-

struct a square loop antenna using segment of straight planar track. The inductance of

each planar track will then be computed using the function shown in Appendix E.2.

Positive and negative mutual inductances between each track are computed by func-

tions shown in Appendix E.3 and Appendix E.3 respectively. The total inductance is

computed by adding the inductance of each track with the total positive mutual induc-

tances and subtracting the total negative mutual inductances.

Section E.2 calculates the inductance of a track, not including the positive and negative

mutual inductances.

Section E.3 calculates the positive mutual inductance between all the tracks.

Section E.4 calculates the negative mutual inductance between all the tracks.

E.1 Main Code

c l e a r a l l ;

2 %in cm

global temp length ;

4 global d e l t a ;

Page 297

E.1 Main Code

global length ;

6 global length1 ;

8 width = 0 . 1 ;

gap = 0 . 0 2 ;

10 length =1;

length1 =2;

12 t h i c k = 0 . 0 0 1 8 ;

n=1;

14 d e l t a =gap+width ;

16 % c r e a t e c o i l

Count=n∗4+1;

18

% Setup

20 S t a r t L o c =[0 , length , 0 ] ;

Curr Loc= S t a r t L o c ;

22 Ct =1;

24 while ( t rue )

26 i f Ct==1

%1 s t Arm

28 temp length ( Ct ) =length1 ;

d i r e c t i o n ( Ct ) =1;

30 e l s e i f

Ct>=Count

32 %f i n a l arm

temp length ( Ct ) =(n ) ∗width+length1 +(n−1)∗gap ;

34 d i r e c t i o n ( Ct ) =1;

break ;

36 else

%t o p arm

38 temp length ( Ct ) = ( ( Ct−1)/2)∗width+2∗ length1 +( Ct−3)/2∗gap ;

d i r e c t i o n ( Ct ) =1;

40 end

Page 298

Appendix E MATLAB Code for Inductance Calculation

42 Ct=Ct +1;

44 %r i g h t arm

temp length ( Ct ) =( Ct /2)∗width+2∗ length +( Ct/2−1)∗gap ;

46 d i r e c t i o n ( Ct ) =2;

48 Ct=Ct +1;

50 %down arm

temp length ( Ct ) = ( ( Ct−1)/2)∗width+2∗ length1 +( Ct−3)/2∗gap ;

52 d i r e c t i o n ( Ct ) =3;

Ct=Ct +1;

54

%l e f t arm

56 temp length ( Ct ) =( Ct /2)∗width+2∗ length +( Ct/2−1)∗gap ;

d i r e c t i o n ( Ct ) =4;

58

Ct=Ct +1;

60

end

62

%c a l c u l a t e t h i n f i l m i n d u c t a n c e L 0

64

for i =1 : Count

66

t h i n f i l m ( i ) =induc ( temp length ( i ) , width , t h i c k ) ;

68

end

70

L=sum( t h i n f i l m ) ;

72 %Compute ( + ) mutual p a i r

temp ct =1;

74

for i =1 : Count−4

76

j = i ;

78

Page 299

E.1 Main Code

while j <=Count−4

80 j = j +4 ;

p pa i r ( temp ct , 1 ) = i ;

82 p pa i r ( temp ct , 2 ) = j ;

temp ct=temp ct +1;

84

end

86 end

88 %Compute (−) mutual p a i r

temp ct =1; i =Count ; while i >0

90

for j =1 : Count ;

92 i f d i r e c t i o n ( j ) ==3

n pa i r ( temp ct , 1 ) = i ;

94 n pa i r ( temp ct , 2 ) = j ;

temp ct=temp ct +1;

96 end

98 end

i =i −4;

100 end

102 i =Count−3;

104 while i >0

106 for j =1 : Count ;

i f d i r e c t i o n ( j ) ==4

108 n pa i r ( temp ct , 1 ) = i ;

n pa i r ( temp ct , 2 ) = j ;

110 temp ct=temp ct +1;

end

112 end

i =i −4;

114 end

Page 300


116 %compute p p a i r

sum p pair =0;

118

for i =1 : s ize ( p pa i r )

120

arm1=p pair ( i , 1 ) ;

122 arm2=p pair ( i , 2 ) ;

124 sum p pair=sum p pair+mutual pair pos ( arm1 , arm2 ) ;

end

126

128 %compute n p a i r

sum n pair =0;

130

for i =1 : s ize ( n pa i r )

132

arm1=n pai r ( i , 1 ) ;

134 arm2=n pai r ( i , 2 ) ;

sum n pair=sum n pair+mutual pair neg ( arm1 , arm2 ) ;

136 end

138 L ;

2∗ sum n pair ;

140 2∗ sum p pair ;

L t o t a l =L+2∗( sum p pair/1000−sum n pair /1000) ;

E.2 Inductance Calculation

1 %C a l c u l a t e i n d u c t a n c e o f t h i n f i l m i n d u c t o r

%

3 %l e n g t h in cm

%width in cm

5 %t h i c k in cm

%

7

function x = induc ( length , width , t h i c k )

9

Page 301

E.3 Mutual Inductance - Positive

l =length ; w=width ; t = t h i c k ;

11

temp =0.002∗ l ∗ ( log (2∗ l /(w+ t ) ) +0.50049+(w+ t ) /3/ l ) ;

13

x=temp ;

E.3 Mutual Inductance - Positive

%C a l c u l a t e Mutual I n d u c t a n c e f o r a p a i r o f s t r i p s

2 %

%l e n g t h in cm

4 %gmd in cm

%

6

function x = mutual pair pos ( arm1 , arm2 )

8



12 l t h 1 =temp length ( arm1 ) ;

l t h 2 =temp length ( arm2 ) ;

14

%a v e r a g e and compute s i d e d i s t a n c e

16 s i d e d i f f =abs ( l th1−l t h 2 ) /2;

18 %f i n d t h e s h o r t e r l e n g t h

min length=min ( l th1 , l t h 2 ) ;

20

%compute gmd v a l u e

22 gmd temp=abs ( arm1−arm2 ) /4∗d e l t a ;

24 i f s i d e d i f f ==0

x=mutual ( min length , gmd temp ) ;

26 else

x=mutual ( min length+ s i d e d i f f , gmd temp )−mutual ( s i d e d i f f , gmd temp ) ;

28 end

Page 302


E.4 Mutual Inductance - Negative

%C a l c u l a t e N e g a t i v e Mutual I n d u c t a n c e f o r a p a i r o f s t r i p s

2 %

%l e n g t h in cm

4 %gmd in cm

%

6

function x = mutual pair neg ( arm1 , arm2 )

8



global length ;

12 global length1 ;

14 l t h 1 =temp length ( arm1 ) ;

l t h 2 =temp length ( arm2 ) ;

16

%a v e r a g e and compute s i d e d i s t a n c e

18 s i d e d i f f =abs ( l th1−l t h 2 ) /2;

20 %f i n d t h e s h o r t e r l e n g t h

min length=min ( l th1 , l t h 2 ) ;

22

%compute gmd v a l u e

24 gmd temp=abs ( arm1−arm2 ) /4∗d e l t a ;

26 % adding gap t o gmd

% i f odd

28 i f rem ( arm1 , 2 ) ==1

gmd temp=gmd temp+2.∗ length ;

30 else

gmd temp=gmd temp+2.∗ length1 ;

32 end

34 i f s i d e d i f f ==0

x=mutual ( min length , gmd temp ) ;

36 else

Page 303

E.4 Mutual Inductance - Negative

x=mutual ( min length+ s i d e d i f f , gmd temp )−mutual ( s i d e d i f f , gmd temp ) ;

38 end

Page 304

Appendix F

Path Loss Experiment

F.1 Preliminary Setting Up Procedure

1. Each of the two antennas was mounted on a tripod stand as shown in Fig. F.1.

Both antennas were positioned at the same height and the distance from the

ground to the centre point of the antenna, h, was measured. The value obtained

was h = 1.29 m.

Figure F.1. Antennas on tripods.

2. A signal of 10 dBm at a frequency of 902 MHz from the HP ESG-3000A Signal

Generator was fed into one antenna. While 2 W ERP is equivalent to +33 dBm,

10 dBm was a convenient output power, well within the capability of the Signal

Generator. In addition, while the frequency of operation in Europe is from 865

Page 305

F.2 Experiment Procedure and Results

to 868 MHz, the antennas that we had available were rated for operation from

902 to 928 MHz. The lower 902 MHz was chosen so as to be within the operating

parameters of our antennas and be as close to the European frequency band.

3. The spectrum analyser was set to a range of 900 to 904 MHz. The other antenna

was connected to the HP 8594E Spectrum Analyser with a 10 dB attenuator at the

input to the spectrum analyser.


F.2.1 Map

Map showing the locations of measurements:

RFID

Lab

Office

Office

Antenna

(Transmitter)To EM

building

Antenna

(Receiver)

RFID

Lab

Office

Office

Antenna

(Transmitter)To EM

building

Antenna

(Receiver)

(a) (b)

Figure F.2. Measurement of signal strength. (a) within a room, and (b) between rooms in

Engineering North building

F.2.2 Signal Strength at Different Distances (Within and Between

Buildings)

1. With the antennas setup as described above, the second antenna (connected to

the spectrum analyser) was moved around various points in the building, with

propagation through walls and typical building materials. The second antenna

was also moved to other nearby buildings. (Please refer to previous for the path

Page 306

Appendix F Path Loss Experiment

RFID

Lab

Office

Office

ENGINEERING NORTH(West Side)

Path taken when

measuring signal strength

Antenna

(Transmitter)

To Maths

building

To EM

building

To Engineering South

Figure F.3. Path taken when measuring signal strength from the west side of Engineering

North building.

RFID

Lab

Office

Office

Antenna

(Transmitter)

Path taken when

measuring signal strength

To EM

building

ENGINEERING NORTH(East Side)

Figure F.4. Path taken when measuring signal strength from the east side of Engineering

North building.

RFID

LabAntenna

(Transmitter)

From

Engineering

North

Path taken when measuring

signal strength

ENGINEERING & MATHS (EM)BUILDING

Figure F.5. Path taken when measuring signal strength from Engineering and Maths (EM)

building.

Page 307


taken when measuring the signal strength.) For all measurements, the orientation

of Antenna 1 was adjusted so as to point approximately towards the expected

location of Antenna 2, see map in previous section.

2. The strength of the received signal was recorded and the straight-line distance

from the transmitting antenna to the receiving antenna was approximated at each

point.

3. The free space path loss was calculated using the equation, Ploss = 20 log10

(4πd

l

),

where d is the distance between the transmitting and receiving antennas and l

is calculated to be 0.333 m for a frequency of 902 MHz. Considering a signal

of amplitude 10 dBm, antennas of gain 8 dBi each, cable loss of 1.5 dBm (on

both the transmitter and the receiver sides respectively), a 10 dB attenuation from

the attenuator, and using the free space path loss obtained, the expected signal

strength at various points was calculated.

4. The measured signal strengths obtained from various points were then compared

with their respective calculated (expected) signal strength. The results are shown

in Table F.1 below. The graph of measured and calculated signal strengths against

distance, d, was plotted and is as shown below in Fig. F.6.

Comments on results: From the results in Table F.1, it can be observed that the mea-

sured strength of the received signal was stronger on the east side of the Engineering

North building as compared to the west side. The reason a weaker signal strength was

detected on the west side may be due to the reason that an anechoic chamber is located

in between the propagation path of the signal. It can also be observed from the results

that, when measurement was done in the east end of the Engineering and Maths (EM)

building, the strength of the received signal was slightly better compared to the mea-

surement taken in front of the lift in that same building. This is because, in the east

end, there was an open space in the middle of the signal propagation path, and lots of

glass windows, as opposed to brick walls.

Page 308


Table F.1. Measured and calculated signal strengths for various locations. d is the distance

between antennas; Pcal is the calculated received power; Pmea is the measured received

power.

d (m) Pcal (dBm) Pmea (dBm) Location/Comments

8.25 -37 -45 Within room (N203)

4 -31 -35 Within room (N203)

18 -44 -70 Between rooms (N203 - N220b)

36 -40 -75 West of engineering north building. Attenuator

pad removed.

51 -43 -85 Further west of engineering north building.

52 -43 -90 Further west of engineering north building. Fre-

quency span of spectrum analyser reduced to

901.9 to 902.1 MHz.**



25 -37 -63 East of engineering north building.

33 -39 -60 Further east of engineering north building.

44 -41 -74 Further east of engineering north building.

28 -37 -78 At north of EM building (3rd Floor, in front of

lift)

37.5 -40 -71 At north of EM building (3rd Floor, east end,

propagation through glass and open area)

28 -37 -90 At north of EM building (4th Floor, in front of

lift, propagation through floor)

37.6 -40 -85 At north of EM building (4th Floor, east end,

propagation through glass, open area and floor)

Page 309


Comparison Between Calculated and Measured Received

Signal Strength In Building

-31

-37

-44-46-45

-43-43-41

-40-40-40-39-37-37-37

-90

-100

-96

-90

-85

-74

-85

-71

-75

-60

-78

-63

-70

-45

-35

-120

-100

-80

-60

-40

-20

0

0 10 20 30 40 50 60 70 80

Distance of Separation Between Antennas, d (m)

Calculated

Measured

Straight line approximations

Re

ce

ive

dS

ign

al

Str

en

gth

(dB

m)

Figure F.6. Comparison between calculated and measured received signal strength (in build-

ing).

F.2.3 Reflection from Typical Wall and a Conductive Fence

1. The antennas were setup according to the description in Part (A) above. The

settings for the signal generator remained the same, and the spectrum analyser

was set to a frequency range of 901.9 to 902.1 MHz with the 10 dB attenuator at

the input removed.

2. Both antennas were arranged and positioned to face the wall (Fig. F.7) and the

strength of the received signal was measured. It has to be noted that the antennas

should not be placed too close to each other to avoid cross coupling.

3. Step (2) was repeated except that this time, both the antennas were positioned to

face to the same point on the wall (Fig. F.8).

4. Step (2) was again repeated, but this time with the antennas facing a conductive

fence instead of the wall (Fig. F.9).

Page 310


WA

LL

Transmitting

Antenna

Receiving

Antenna

(a) (b)

Figure F.7. Antennas facing directly perpendicular to the wall.

WA

LL

Transmitting

Antenna

Receiving

Antenna

(a) (b)

Figure F.8. Antennas facing with an angle to the wall.

Figure F.9. Antennas on tripods facing a conductive fence.

5. The results are as shown in Table F.2 below.

Page 311


Table F.2. Measured signal strength for antennas facing a wall or a conductive fence. Ds is

the straight line distance between antennas; Dp is the perpendicular distance from wall

(or fence); Prec,d is the measured received power when antennas are pointing directly to

wall; Prec,0 is the measured received power when antennas are pointing at same point on

wall.

Ds Dp Prec,d Prec,0

Antennas facing wall (1st) 2.5 2.3 -41 -45

Antennas facing wall (2nd) 2.2 1.5 -45 -38

Antennas facing conductive fence 1.77 1.66 -32 —

F.2.4 Propagation Loss Outdoors

1. This part of the experiment was performed outdoor, on the lawn in front of the

Mathematics building.

2. The antennas were setup in a similar way. The settings for the signal generator

remained the same, and the spectrum analyser was set to a frequency range of

901.9 to 902.1 MHz with the 10 dB attenuator at the input removed.

3. The antennas were arranged and positioned to face each other (Fig. F.10) and the

distance of separation between the two antennas, d, was measured. It was found

that d = 11 m. The height of the antennas from the ground, h, was also measured

and it was found to be 1.29 m.

4. The strength of the received signal was measured.

5. The procedures were repeated with both antennas placed on the ground and fac-

ing upwards (with the distance, d, maintained the same) (Fig. F.11).

6. The recorded results are as shown in Table F.3 below.

Page 312


(a) (b)

Figure F.10. Measurement of signal strength in outdoors.

Figure F.11. Antennas on the ground and facing upwards.

F.3 Conclusion

As expected, the measured values of the received signal strength were found to be

quite different from the calculated values, as obtained from the Free Space Loss Equa-

tion. Fortunately, the observed path loss was greater than the path loss calculated in

free space, due to building materials and objects in various rooms.

Our transmitter operated at +10 dBm, whereas RFID transmitters are allowed to oper-

ate at 2 W ERP, or 35 dBm. The path loss required would then be 25 dB greater than

our measured results, i.e.121 dBm (96 + 25). Using Fig. F.6, a distance of at least 110 m

would be required for the separation of two readers, where one is already operating

in a particular sub-band and the other listening before wanting to talk in the same

sub-band.

Page 313

F.3 Conclusion

Table F.3. Calculated and measured signal strength (outdoor) for different antenna orien-

tations. d is the distance between antennas; Prec is calculated received power; Pmea is

the measured received power.

Antenna orientation Distance d (m) Prec (dBm) Pmea (dBm)

Antennas facing directly at each other 11 -29 -39

Antennas on the ground, facing upwards 11 — -71

Page 314

Appendix G

RFID Deployment inPiggery

This appendix presents all the design steps for RFID deployment in a piggery. In the

first part the modelling of a pig feeder in laboratory is shown. In the second part, the

final report on the RFID deployment in piggery is included, which covers the HF and

UHF tags designed, reader installation at piggery, experiment data and analysis, and

conclusion.

G.1 Modelling of Pig Feeder

Raw Material:

1. 2 × Zinc plate (600mm × 1000mm)

2. 1 × Zinc plate (600mm × 400mm)

3. 2 × Wooden piece (400mm × 900mm)

4. Small Zinc plates

5. Bolts and nuts

Note: The dimensions of Zinc plate (Item 1 and 2) are increased by 50 mm on each side,

so that the extra length can be bent and bolted onto.

Construction of Main Body:

Page 315

G.1 Modelling of Pig Feeder

(a)

A

B

CC

DE

F

G

(b)

Figure G.1. A pig feeder used in RFID deployment experiment. (a) Picture of a pig feeder.

(b) Major parts of a pig feeder to be modelled.

a. Raw materials item 1 (2 × Zinc plates) are bent 50 mm each on 4 sides. The

biggest surface is measured 500 mm × 900 mm.

b. Item 1 (2 × Zinc Plates) are bolted at location A and B onto the two wooden

pieces (Item 3). Refer to Fig. G.1(b) and Fig. G.2 for clarity.

Construction of Feeder Part:

a. Raw material item 2 is bent 50 mm each on 4 sides. The biggest surface is mea-

sured 500 mm x 300 mm. It is bolted in location E.

b. A small Zinc plate is bolted in location D. Refer to Fig. G.1(b) and Fig. G.3 for

clarity. The size is a rough approximation as it depends the on the accuracy of

construction of part A, C and E.

c. A small Zinc plate is bolted in location F. Refer to Fig. G.1(b) and Fig. G.3 for

clarity. The dimension measurement is dependent on A, D, and E.

d. At position G of both sides of the feeder (dotted line), place a Zinc strip (width

of 50 mm). Refer to Fig. G.1(b) and Fig. G.3 for clarity.

Page 316

Appendix G RFID Deployment in Piggery

400

900

90

0

50

70

23

A

B

C

Figure G.2. Modelling of a pig feeder (Stage 1).

G.2 Final Report

This section contains the final report sent to industrial partners based on the field trials

carried out a piggery. This report is slightly modified to avoid repetition. The pictures

appear in the original report has been included in the main body of the thesis and will

be referred to.

G.2.1 Executive Summary

This is a report on the design, manufacture and trial testing of RFID tags for possible

future deployment in a piggery. Two types of tags were produced, employing different

RFID technology, one at HF (13.56 MHz) and the other at UHF (in Australia 920 to

926 MHz). Section G.2.2 presents the design, fabrication and encapsulation of the HF

and UHF RFID tags. Section G.2.3 shows the feeder design and the feeder setup in the

piggery. The Experiment setup is examined in Section G.2.4 with experimental results

Page 317

G.2 Final Report

90

300

315

735

100

D E

F

G

Figure G.3. Modelling of a pig feeder (Stage 2).

discussed in Section G.2.5, followed by conclusions and recommendation for future

studies at the end of this document.

G.2.2 Design of RFID Tag Antennas

This section shows the final design of the RFID tag antenna. RFID tags suitable for

attaching to a pig’s ear must be encapsulated in a strong casing before deployment to

protect the tag antennas from being damaged by the pigs. Hence the design of the tag

antenna is restricted by the size of the casing. As the objective of the project was to

compare the performance of a number of RFID technologies, standard commercially

available Low Frequency (LF) technology tags as marketed by Leader Products were

deployed, and compared with High Frequency (HF) technology and Ultra High Fre-

quency (UHF) technology tags, designed by the Auto-ID Lab, Adelaide. The antennas

on the Printed Circuit Board (PCB) are manufactured by Entech Electronic. For small

Page 318


scale testing, RFID tag chips (both HF and UHF from Texas Instruments) were man-

ually attached onto custom designed RFID tag antennas using Z-axis conductive tape

(Electrically Conductive Adhesive Transfer Tape 9703 from 3M).

1. HF RFID Tag Antenna Design

The final design of the HF antenna is shown in Fig. 6.40. The thickness of the sub-

strate is 0.2 mm, and the coil track separation is 0.3 mm. The chip that used was

the Texas Instrument Tag-it HF-I Standard Transponder IC, supplied by Electro-

Com, which is based on the ISO/IEC 15693 and ISO/IEC 18000-3 Mode 1 stan-

dards, and has a nominal input capacitance of 23.5 pF. The RFID reader used for

identifying HF RFID tags is ID ISC.LR2000 from Feig Electronic.

2. UHF RFID Tag Antenna Design

The final design of the UHF antenna is based on the antenna shown in Fig. 6.49(a).

The thickness of the substrate is 0.2 mm. The antenna is designed to have a

matching impedance with respect to the RFID chip used, which is RI-UHF-STRAP-

08 UHF Gen2 Strap from Texas Instruments, supplied by Electro-Com. The RFID

reader used for identifying UHF RFID tags is ID ISC.LRU2000 from Feig Elec-

tronic.

3. Tag Encapsulation

Tags were encapsulated by Leader Products. Encapsulated tags are shown in

Fig. 6.42(b). Read range performances for both the HF and UHF tags before and

after encapsulation were recorded. It is expected that the encapsulation process

will have a relatively bigger impact on the UHF tags when compared to the HF

tags.

4. Performance of HF RFID Tags

Table G.1 shows the read range performance of HF tags before and after encap-

sulation. Also, after encapsulation, tags are positioned on a human hand to sim-

ulate the actual field trial condition, where tags are attached on the ears of pigs.

As expected, the encapsulation did not affect the tag performance by much. The

performance of tags fluctuates slightly as the fine-tuning was carried manually.

Page 319

G.2 Final Report

Table G.1. Performance of HF RFID tag before and after encapsulation.

Tag No. Free Space (m)After Encapsulation

Free Space (m) On Hand (m)

HF01 0.32 0.32 0.33

HF02 0.18 0.27 0.18

HF03 0.34 0.33 0.25

HF04 0.33 0.32 0.33

HF05 0.32 0.28 0.25

HF06 0.34 0.32 0.35

HF07 0.32 0.31 0.33

HF08 0.25 0.23 0.20

HF09 0.32 0.28 0.34

HF10 0.24 0.26 0.28

HF11 0.29 0.30 0.26

HF12 0.31 0.30 0.30

HF13 0.33 0.33 0.34

HF14 0.26 0.26 0.32

HF15 0.32 0.31 0.35

HF16 0.32 0.31 0.32

HF17 0.33 0.33 0.33

HF18 0.31 0.30 0.33

HF19 0.32 0.30 0.33

HF20 0.33 0.31 0.34

HF21 0.34 0.33 0.25

HF22 0.34 0.34 0.34

HF23 0.34 0.33 0.36

HF24 0.32 0.32 0.26

HF25 0.33 0.32 0.34

HF26 0.33 0.32 0.33

5. Performance of UHF RFID Tags

Similar to Table G.1, Table G.2 shows the read range performance of UHF tags

before and after encapsulation. Also, after encapsulation, tags are positioned on

a human hand to simulate the actual field trial condition, where tags are attached

on the ears of pigs.

Page 320


Table G.2. Performance of UHF RFID tag before and after encapsulation.

Tag No. Free Space (m)After Encapsulation

Free Space (m) On Hand (m)

C1 0.60 0.80 0.34

C2 0.49 1.10 0.33

C3 0.45 0.40 0.30

C4 1.04 1.05 0.32

C5 0.46 0.70 0.35

C6 0.62 0.85 0.30

C7 0.36 0.75 0.34

C8 0.40 0.82 0.43

C9 0.89 0.43 0.40

C10 0.49 0.87 0.36

D1 0.68 0.85 0.36

D2 0.49 2.06 0.38

D3 0.41 1.15 0.40

D4 0.37 2.08 0.40

D5 0.85 0.80 0.47

D6 0.43 0.50 0.30

D7 0.65 0.80 0.31

D8 0.50 0.53 0.33

D9 0.60 0.83 0.37

D10 1.00 0.67 0.35

E1 0.55 0.36 0.33

E2 0.34 0.43 0.24

E3 0.86 0.36 0.31

E4 0.54 0.45 0.29

E5 0.36 1.09 0.30

E6 0.65 0.77 0.29

E7 0.59 1.05 0.40

E8 0.58 0.34 0.29

E9 0.49 0.78 0.35

E10 0.21 0.42 0.30

Page 321

G.2 Final Report

The inconsistency in results is caused by several reasons: (i) All tags are fine-

tuned by hand. (ii) All tags are tuned higher than the operating frequency as

experience shows that encapsulation will de-tune the antenna by reducing the

tag’s resonant frequency.

G.2.3 Feeder Design and Experiment Setup

The trials were conducted at The Department of Primary Industries, Victoria at the Pig

Research & Training Centre, 600 Sneydes Rd, Werribee 3030. Fig. G.1 shows a typical

feeder used at that facility. The objective of the trial is to record when a pig, which has

been tagged with an RFID tag, goes to the feeder to eat. Fig. G.4 is the front view of the

planned feeder with RFID reading capability. Fig. G.5 shows the location of antennas

in the protective casing at both sides of a feeder. The slope of a feeder is made of metal.

The antennas positioned at the sides should not intersect with the metallic slope, as

this will degrade the performance of the RFID tag reading.

The planning for HF and UHF readers in plastic protective casings are shown in Fig. G.6

and Fig. G.7. HF system requires an external multiplexer and an additional power

supply. Fig. 6.44 and Fig. 6.45 are photos taken at the test site, with HF and UHF RFID

system setup and reader antennas embedded in thick plastic protective casing besides

the feeders. UMD was contracted for packaging and installation of the readers and

antennas.

G.2.4 Network Setup

Remote access is set up so that the computer (A personal computer running on Pen-

tium 4 with Windows XP) managing the HF and UHF readers can be controlled from

Adelaide through the Internet. An illustration on how the data retriever system works

is as shown in Fig. G.8.

1. Window Scheduler

Triggers sendlog HF.bat and sendlog UHF.bat everyday 0:30 am.

Page 322


feeder

Change to Non-Metalic Sides

UHF Antenna Box

Reader Box

Cables

Figure G.4. Overview arrangement of RFID deployment in piggery.

2. sendlog HF.bat

Compresses the HF data file one day before, packages the data as an email, and

sends the file to smtp.gmail.com through stunnel.

3. sendlog UHF.bat

Compresses the UHF data file one day before, packages the data as an email, and

sends the file to smtp.gmail.com through stunnel.

4. stunnel

stunnel is a software package used to manage the security of data transmission

using Secure Sockets Layer (SSL) protocol.

Window Scheduler is built-in function of Windows XP. sendlog HF.bat and sendlog UHF.bat

are modified version of software supplied by Behnam Jamali. stunnel is a freeware

available on the Internet.

Page 323

G.2 Final Report

Reader Box

(Side View)

Cable

Slope ofthe feeder

Feeder

l

Antenna Box(Grey Zone)

Figure G.5. The position of an antenna in an antenna box besides a pig feeder.

G.2.5 Experiment Results

The target trial pigs were tagged on 24th September, 2007. Testing was carried from

25th to 27th September, 2007. The times the pigs moved to the feeder to eat were

recorded, so that later the data file could be checked for misreads and ghost reads,

where misreads occur when a tag enters interrogation zone but is not picked up by

reader, and ghost reads refer to the situation when unknown tag numbers appeared in

the log file.

The data files, recording pigs at the feeders were monitored remotely after 27th Septem-

ber, 2007. It is not possible to check for misread, but it is possible to examine the eating

behaviour of individual pigs. If a pig looses its appetite drastically, as evidenced by no

reads on the data file, either the tag has failed, or it has been lost, or the pig may need

medical attention.

Page 324


8080

80

340

330

180

40 40

20

100

a

b

c c

d d e

f f

110

640

20

Figure G.6. Encapsulation of a HF reader.

As shown in Fig. 6.46, when facing a pig, the HF tag is on the left (the pig’s right ear)

while the UHF tag is on the right (the pig’s left ear). When a pig is feeding, as shown

in Fig. 6.47, the HF tag is on the right and the UHF tag is on the left when viewed from

the back of a pig.

1. Initial Test Run

Tags of reasonable performance are randomly picked and attached onto the ears

of pigs. The chosen tags are recorded in Table G.3.

The results from 25th to 27th September 2007 were recorded where:

Page 325

G.2 Final Report

100

320

180

40

20

ab

c d

20

e e

80

440

280

20

Figure G.7. Encapsulation of a UHF reader.

Stunnel

WindowsScheduler

Sendlog_HF.bat

Smtp.gmail.com

Backto Adelaide

Sendlog_UHF.bat

Figure G.8. Network setup in a piggery. Data is automatically sent back to the Adelaide labora-

tory.

• Misread: Tag enter interrogation zone but not picked up by reader.

• Ghost Read: Unknown tag number appeared in the log file.

• Read outside of read zone: Tag not in the designated zone but picked up by

the reader.

Page 326


Table G.3. HF and UHF RFID tags attached to pigs.

Pig HF UHF

1 HF17 C1

2 HF07 D3

3 HF20 D8

4 HF14 D7

5 HF22 D1

6 HF19 C6

7 HF16 D10

8 HF15 E7

9 HF04 C4

10 HF06 C10

Date: 25 Sept. 2007

Time: 1125-1145

HF Misread: 0 misread

HF Ghost Read: 0

HF Read outside of read zone: 1

HF Special Notes: Good performance so far. However, pigs are relatively inac-

tive. More observation needed.

UHF Misread: 0 misread

UHF Ghost Read: 0

UHF Read outside of read zone: 0

UHF Special Notes: Good performance so far. However, pigs are relatively inac-

tive. More observation needed.

Date: 25 Sept. 2007

Time: 1347-1427


HF Ghost Read: 0



Page 327

G.2 Final Report

UHF Ghost Read: 1 (426; EPC Gen 2; 00065090C600EE000C06080000; 14:55:09.834

09/25/2007;|1|;)UHF Read outside of read zone: 0

Date: 26 Sept. 2007

Time: 0931-1034


HF Ghost Read: 0


HF Special Notes: High Read outside of read zone were caused by active pigs

running near the side of the RHS antenna. Pigs running near the side of the RHS

antenna are manually recorded and coincide with the pig recorded by the reader.

This shows an obstacle can efficiently reduce the Read outside of read zone.

Recommendation: Put obstacles near the side of the feeder to avoid pig getting

into the unintended read zone.


UHF Ghost Read: 0


Date: 27 Sept. 2007

Time: 1315-1345

HF Misread: 3 misread within 30 minutes.

HF Ghost Read: 0

HF Read outside of read zone: Several occasions when a pig gets very near to the

outer RHS antenna.

HF Special Notes:LHS Antenna not performing well. This is caused by the nearby

metal sheet. Furthermore, tags are tagged on the left hand side of the pig if

viewed from front (Refer Fig. 6.46), making the detection of the pig on the left

even harder.


Page 328


UHF Ghost Read: 0


UHF Special Notes: At 1331, Pig 0 jumped and landed on pig 7, which is eating

at the feeder. The reader picked up a single read as Pig 0 extended its head into

the read zone.

2. Distance Monitoring

Data was collected remotely to monitor the eating behaviour of tagged pigs. The

total read count fluctuates due to power failures in the computer room at the

test site. The results are as shown in Table G.4, where the numbers of reads for

each tag (both HF and UHF) are shown. A pig feeding at HF feeder will only be

recorded by the HF reader, not by the UHF reader, and vice versa.

A few observations: (1) UHF tag 8 failed on 30 Sept., 7 days after it was tagged

on the pig ear. It was discovered later that this tag has dropped off the pig ear.

(2) It can be seen that both Pig 6 and Pig 8 are relatively inactive. (3) The other

pigs appear to be eating normally.

G.2.6 Conclusions

Based on the initial test run, which spans 25th to 27th September 2007 with manual

observation:

1. The performance of RFID UHF system is excellent, with no misreads. The only

error recorded is a single ghost read on 25th Sept. 2007. The reason for the ghost

read is unknown. According to the reader manufacturer, FEIG, ghost reads are

extremely rare.

2. However, the performance of the RFID HF system is not as good. Misreads oc-

cur in almost every session. Misread always happens on the left hand side of

the feeder. The reason can be (i) The left hand side antenna is near to a UHF

reader antenna. Although arrangement has been made to shift the antenna fur-

ther away, it is suspected a greater distance is required. (ii) The HF tag is attached

Page 329

G.2 Final Report

Table G.4. Results from remote monitoring.

Date Pig 1 2 3 4 5 6 7 8 9 0 Read

27 Sept.HF 601 348 305 330 569 425 280 38 116 206 3218

UHF 325 489 311 478 316 56 336 159 533 628 3631

28 Sept.HF 344 254 221 146 471 100 153 351 124 287 2451

UHF 406 293 189 296 369 59 304 88 414 411 2829

29 Sept.HF 523 96 289 77 616 123 404 55 11 222 2416

UHF 470 528 352 417 449 143 320 92 314 612 3697

30 Sept.HF 215 38 83 57 319 112 114 20 63 70 1091

UHF 139 206 28 130 65 56 56 2 190 87 959

1 Oct.HF 432 351 394 196 641 233 108 122 88 342 2907

UHF 107 245 63 118 99 59 302 0 93 129 1215

2 Oct.HF 336 329 383 159 544 182 276 73 67 237 2586

UHF 276 502 152 398 175 177 482 0 597 393 3152

3 Oct.HF 88 83 64 32 403 53 125 47 32 48 975

UHF 488 382 297 370 255 196 569 0 519 558 3634

4 Oct.HF 457 651 468 171 573 251 424 131 232 882 4240

UHF 605 305 405 450 469 190 796 0 561 740 4521

to the right ear of a pig. When a pig is eating at the feeder, the HF tag will be ap-

proximately 200 mm from the left antenna and 400 mm from the right antenna.

The left antenna should pick up the tag, but due to the degradation as mentioned

in (i), a misread can occur.

3. The tag encapsulation from Leader Product is excellent. So far all tags are still

functioning properly, and not damaged by pigs through constant chewing.

4. Installations of HF and UHF readers by UMD are of good standard.

Based on the distance monitoring:

1. UHF tag 8 stopped giving read data from 30 September onwards. It was discov-

ered later that UHF tag 8 dropped off from the pig’s ear.

Page 330


Recommendation for Future Studies:

While this trial appears to indicate that the UHF system is more reliable and robust,

and for the foreseeable future, UHF tags are likely to be cheaper that HF tags, the HF

system should not be ruled out. In a future trial it is recommended that more fine-

tuning may be necessary, either to provide more space around the feeders, to reduce

the number of misreads and reads outside the zone, or as suggested by Bruce Dumbrell

from Leader Product, the HF antenna can be put on top or under the feeder entry point,

rather than the current position which is at the sides of the feeder entry point.

Page 331

Bibliography

∗Please refer to the Conventions Section (page xxi) for descriptions of the citation style used inthis bibliography.

[1] M. Shell. (2007, 11 Jan.) How to Use the IEEEtran LATEXClass. [Online]. Available: http://tug.

ctan.org/tex-archive/macros/latex/contrib/IEEEtran/IEEEtran HOWTO.pdf [12 Nov. 2007].

[2] M. Shell. (2007, 11 Jan.) The IEEEtran BIBTEX style. [Online]. Available: http://www.ctan.org/

tex-archive/macros/latex/contrib/IEEEtran/bibtex/IEEEtran bst HOWTO.pdf [12 Nov. 2007].

[3] Bureau International des Poids et Mesures. (2007) The international system of units (SI). [Online].

Available: http://www.bipm.org/utils/common/pdf/si brochure 8 en.pdf [12 Nov. 2007].

[4] EPCglobal. (2007, 28 Sep.) EPCglobal Tag Data Standards version 1.3.1. [Online]. Available: http:

//www.epcglobalinc.org/standards technology/EPC TagDataSpecification11Rev126ca.pdf [12

Nov. 2007].

[5] Institute of Electrical and Electronics Engineers. (2007) IEEE technical societies/councils. [Online].

Available: http://www.ieee.org/portal/pages/tab/society.html [12 Nov. 2007].

[6] J. Landt, “The history of RFID,” Potentials, IEEE, vol. 24, no. 4, pp. 8–11, 2005.

[7] D. C. Ranasinghe, K. S. Leong, M. L. Ng, D. W. Engels, and P. H. Cole, “A distributed architec-

ture for a ubiquitous item identification network,” in Seventh International Conference on Ubiquitous

computing, Tokyo, Japan, 11-14 Sept. 2005.

[8] D. C. Ranasinghe, K. S. Leong, M. L. Ng, D. W. Engels, and P. H. Cole, “A distributed architecture

for a ubiquitous RFID sensing network,” in 2nd International Conference on Intelligent Sensors, Sensor

Networks and Information Processing (ISSNIP), Melbourne, Australia, 5-8 Dec. 2005.

[9] J. Landt, and B. Catlin, “Shrouds of time, the history of RFID,” AIM Publication, 2001.

[10] D. B. Harris, and P. Alto, “Radio transmission systems with modulatable passive responder,” U.S.

Patent Office, Patent 2,927,321, 1960.

[11] R. F. Harrington, “Small resonant scatterers and their use for field measurements,” Microwave

Theory and Techniques, IEEE Transactions on, vol. 10, no. 3, pp. 165–174, 1962.

[12] R. Harrington, “Electromagnetic scattering by antennas,” Antennas and Propagation, IEEE Transac-

tions on [legacy, pre - 1988], vol. 11, no. 5, pp. 595–596, 1963.

[13] P. H. Cole, “Electronic surveillance system,” Australian Patent No. 444,838, U.S. Patent No.

3,706,094, 1972.

Page 333

Bibliography

[14] A. R. Koelle, S. W. Depp, and R. W. Freyman, “Short-range radio-telemetry for electronic identi-

fication, using modulated RF backscatter,” Proceedings of the IEEE, vol. 63, no. 8, pp. 1260–1261,

1975.

[15] K. Finkenzeller, RFID handbook, 2nd ed. John Wiley and Sons, 2003.

[16] International Standards Organization, “ISO/IEC 19762-3: information technology - automatic

identification and data capture (AIDC) techniques - harmonized vocabulary - part 3: radio fre-

quency identification (RFID),” 2005.

[17] U. Karthaus, and M. Fisher, “Fully integrated passive UHF RFID transponder IC with 16.7µ W

minimum RF input power,” Solid-State Circuits, IEEE Journal of, pp. 1602 – 1608, 2003.

[18] R. Glidden, C. Bockorick, S. Cooper, C. Diorio, D. Dressler, V. Gutnik, C. Hagen, D. Hara, T. Hass,

T. Humes, J. Hyde, R. Oliver, O. Onen, A. Pesavento, K. Sundstrom, and M. Thomas, “Design

of ultra-low-cost UHF RFID tags for supply chain applications,” Communications Magazine, IEEE,

vol. 42, no. 8, pp. 140–151, 2004.

[19] S. Lewis. (2004, Jan.) A basic introduction to RFID technology and its use in the supply

chain. [Online]. Available: http://www.printronix.com/library/assets/public/case-studies/

rfid-laran-white-paper-english.pdf [12 Nov. 2007].

[20] G. S. Pope, M. Y. Loukine, D. M. Hall, and P. H. Cole, “Innovative systems design for 13.56 MHz

RFID,” in Wireless and Portable Design Conference, Burlington, Massachusetts, 1997, pp. 240–245.

[21] Philips Semiconductor, TAGSYS, and Texas Instruments Inc. (2004, July 2004) Item-level

visibility in the pharmaceutical supply chain: a comparison of HF and UHF RFID

technology. [Online]. Available: http://www.tagsysrfid.com/modules/tagsys/upload/news/

TAGSYS-TI-Philips-White-Paper.pdf [12 Nov. 2007].

[22] R. Want, “RFID - a key to automating everything,” Scientific American, vol. 290, no. 1, pp. 56–65,

2004.

[23] G. Goth, “RFID: not quite prime time, but dawdle at your own risk,” IEEE Distributed Systems

Online, vol. 6, no. 2, 2005.

[24] C. R. Nave. (2007) Transparency of water in the visible range. [Online]. Available:

http://hyperphysics.phy-astr.gsu.edu/hbase/chemical/watabs.html [12 Nov. 2007].

[25] P. H. Cole, and D. W. Engels, “Auto-ID - 21st century supply chain technology,” in AEEMA Na-

tional Conference, Melbourne Exhibition Centre, Melbourne, Australia, 2002.

[26] R. Angeles, “RFID technologies: supply-chain applications and implementation issues,” Informa-

tion Systems Management, vol. 22, no. 1, pp. 51–65, 2005.

Page 334

Bibliography

[27] T. J. Becker. (2000, 12 Dec.) Putting a price on supply chain problems: study links supply chain

glitches with falling stock prices. [Online]. Available: http://www.gtresearchnews.gatech.edu/

newsrelease/CHAINR.html [12 Nov. 2007].

[28] M. N. Huhns, and L. M. Stephens, “Automating supply chains,” Internet Computing, IEEE, vol. 5,

no. 4, pp. 90–93, 2001.

[29] D. Luckett, “The supply chain,” BT Technology Journal, vol. 22, no. 3, pp. 50–55, 2004.

[30] K. M. Penttila, D. W. Engels, and M. A. Kivikoski, “Radio frequency identification systems in

supply chain management,” International Journal of Robotics & Automation, vol. 19, no. 3, pp. 143–

151, 2004.

[31] R. Weinstein, “RFID: a technical overview and its application to the enterprise,” IT Professional,

vol. 7, no. 3, pp. 27–33, 2005.

[32] Auto-ID Lab. (2007) History of Auto-ID Labs. [Online]. Available: http://autoidlab.eleceng.

adelaide.edu.au/ [12 Nov. 2007].

[33] International Standards Organization, “ISO/IEC 18000-3:2004: information technology - radio fre-

quency identification for item management - Part 3: parameters for air interface communications

at 13.56 MHz,” 2004.

[34] International Standards Organization, “ISO/IEC 15693-1:2000: identification cards - contactless

integrated circuit(s) cards - vicinity cards (Part 1 - 3),” 2000-2001.

[35] EPCglobal. (2007, 17 Oct.) EPC radio-frequency identity protocols class-1 generation-2 UHF

RFID protocol for communications at 860 MHz - 960 MHz version 1.1.0. [Online]. Available:

http://www.epcglobalinc.org/standards/uhfc1g2 [12 Nov. 2007].

[36] D. Barlas. (2003, 4 June) Wal-Mart’s RFID mandate. [Online]. Available: http://www.line56.

com/articles/default.asp?ArticleID=4710 [1 Dec. 2006].

[37] Department of Defence United States of America. (2007, 26 Oct.) Radio frequency identification

(RFID). [Online]. Available: http://www.acq.osd.mil/log/rfid/index.htm [12 Nov. 2007].

[38] S. Sarma, “Integrating RFID,” Queue, vol. 2, no. 7, pp. 50–57, 2004.

[39] U.S. Food and Drug Administration. (2004, 15 Nov.) FDA announces new initiative to protect

the U.S. drug supply through the use of radiofrequency identification technology. [Online].

Available: http://www.fda.gov/bbs/topics/news/2004/NEW01133.html [12 Nov. 2007].

[40] S. Sarma. (2001, 1 Nov.) Towards the 5 cents tag. [Online]. Available: http://www.autoidlabs.

org/uploads/media/mit-autoid-wh-006.pdf [12 Nov. 2007].

[41] EPCglobal. (2007, 10 Sep.) The EPCglobal Architecture Framework version 1.2. [Online].

Available: http://www.epcglobalinc.org/standards/architecture/ [12 Nov. 2007].

Page 335

Bibliography

[42] EPCglobal. (2005, 15 Sep.) The Application Level Events (ALE) Specification version 1.0. [Online].

Available: http://www.epcglobalinc.org/standards/ale [12 Nov. 2007].

[43] EPCglobal. (2007, 21 Sep.) EPC Information Services (EPCIS) version 1.0.1 specification. [Online].

Available: http://www.epcglobalinc.org/standards/epcis [12 Nov. 2007].

[44] EPCglobal. (2005, 4 Oct.) Object Naming Service (ONS) version 1.0. [Online]. Available:

http://www.epcglobalinc.org/standards/ons [12 Nov. 2007].

[45] EPCglobal. (2007, 31 May) Reader Management (RM) Standard version 1.0.1. [Online]. Available:

http://www.epcglobalinc.org/standards/rm [12 Nov. 2007].

[46] EPCglobal. (2006, 21 Jun.) Reader Protocol Standard version 1.1. [Online]. Available:

http://www.epcglobalinc.org/standards/rp [12 Nov. 2007].

[47] Federal Communications Commision, “Title 47, telecommunication, chapter 1, Part 15, radio fre-

quency devices,” 2001.

[48] ETSI, “EN 302 208-1 v1.1.1: electromagnetic compatibility and radio spectrum matters (ERM);

radio frequency identification equipment operating in the band 865 MHz to 868 MHz with power

levels up to 2W,” European Standard (Telecommunications Series), 2004.

[49] International Organization for Standardization. (2007) ISO - International Organization for

Standardization. [Online]. Available: http://www.iso.org/iso/en/ISOOnline.frontpage [12 Nov.

2007].

[50] EPCglobal. (2007) EPCglobal homepage. [Online]. Available: http://www.epcglobalinc.org/

home [12 Nov. 2007].

[51] International Standards Organization, “ISO 11784:1996: Radio-frequency identification of animals

- code structure,” 1996.

[52] International Standards Organization, “ISO 11785:1996: Radio-frequency identification of animals

- technical concept,” 1996.

[53] International Standards Organization, “ISO/IEC 18000-6:2004: information technology - radio fre-

quency identification for item management - Part 6: parameters for air interface communications

at 860 MHz to 960 MHz,” 2004.

[54] EPCglobal. (2003, 1 Feb.) 13.56 MHz ISM band class 1 radio frequency identification

tag interface specification: candidate recommendation, version 1.0.0. [Online]. Avail-

able: http://www.epcglobalinc.org/standards/specs/13.56 MHz ISM Band Class 1 RFID Tag

Interface Specification.pdf [12 Nov. 2007].

[55] B. Jamali, “High-performance RFID System,” Ph.D. dissertation, University of Adelaide, 2005.

Page 336

Bibliography

[56] K. S. Leong, M. L. Ng, and P. H. Cole, “The reader collision problem in RFID systems,” in IEEE

2005 International Symposium on Microwave, Antenna, Propagation and EMC Technologies for Wireless

Communications, Beijing, China, 8-12 Aug. 2005.

[57] K. S. Leong, M. L. Ng, and P. H. Cole, “Positioning analysis of multiple antennas in a dense

RFID reader environment,” in 2006 International Symposium on Applications and the Internet (SAINT)

Workshop, RFID and Extended Network: Deployment of Technologies and Applications, Phoenix, Ari-

zona, USA, 23-27 Jan. 2006.

[58] T. S. Rappaport, Wireless communications - principles and practice, 2nd ed. Prentice Hall, 2002.

[59] ETSI, “EN 302 208-1 v1.2.1: electromagnetic compatibility and radio spectrum matters (ERM);

radio frequency identification equipment operating in the band 865 MHz to 868 MHz with power

levels up to 2W,” European Standard (Telecommunications Series), 2007.

[60] D. Kotz, C. Newport, R. S. Gray, J. Liu, Y. Yuan, and C. Elliott, “Experimental evaluation of wire-

less simulation assumptions,” in Proceedings of the 7th ACM international symposium on Modeling,

analysis and simulation of wireless and mobile systems. Venice, Italy: ACM Press, 2004, pp. 78–82.

[61] D. Eckhardt, and P. Steenkiste, “Measurement and analysis of the error characteristics of an in-

building wireless network,” in Conference proceedings on Applications, technologies, architectures, and

protocols for computer communications. Palo Alto, California, United States: ACM Press, 1996, pp.

243–254.

[62] K. Pawlikowski, H.-D. J. Jeong, and J.-S. R. Lee, “On credibility of simulation studies of telecom-

munication networks,” Communications Magazine, IEEE, vol. 40, no. 1, pp. 132–139, 2002.

[63] K. S. Leong, M. L. Ng, and P. H. Cole, “Operational considerations in simulation and deployment

of RFID systems,” in 17th International Zurich Symposium on Electromagnetic Compatibility, Singa-

pore, 27 Feb. - 3 Mar. 2006.

[64] K. S. Leong, M. L. Ng, and P. H. Cole, “Investigation of the threshold of second carrier sensing in

RFID deployment,” in 2006 International Symposium on Applications and the Internet (SAINT) Work-

shop, RFID and Extended Network: Deployment of Technologies and Applications, Hiroshima, Japan,

15-19 Jan. 2007.

[65] Workgroup 4 - Radio Frequency Identification for Item Managment (ISO/IEC JTC 1/SC 31/WG4),

“Regulatory status for using RFID in the UHF spectrum,” 2005.

[66] H. Hashemi, “The indoor radio propagation channel,” Proceedings of the IEEE, vol. 81, no. 7, pp.

943–968, 1993.

[67] G. E. Athanasiadou, and A. R. Nix, “A novel 3-D indoor ray-tracing propagation model: the path

generator and evaluation of narrow-band and wide-band predictions,” Vehicular Technology, IEEE

Transactions on, vol. 49, no. 4, pp. 1152–1168, 2000.

Page 337

Bibliography

[68] R. Valenzuela, “A ray tracing approach to predicting indoor wireless transmission,” in Vehicular

Technology Conference, 1993 IEEE 43rd, 1993, pp. 214–218.

[69] S. Y. Seidel, and T. S. Rappaport, “914 MHz path loss prediction models for indoor wireless com-

munications in multifloored buildings,” Antennas and Propagation, IEEE Transactions on, vol. 40,

no. 2, pp. 207–217, 1992.

[70] J.-F. LaFortune, and M. Lecours, “Measurement and modeling of propagation losses in a building

at 900 MHz,” Vehicular Technology, IEEE Transactions on, vol. 39, no. 2, pp. 101–108, 1990.

[71] J. Mitsugi, “Two stage carrier sensing to protect communication links in UHF RFID system,” to be

published, 2006.

[72] K. S. Leong, M. L. Ng, A. Grasso, and P. H. Cole, “Synchronisation of RFID readers for dense RFID

reader environments,” in 2006 International Symposium on Applications and the Internet (SAINT)

Workshop, RFID and Extended Network: Deployment of Technologies and Applications, Phoenix, Ari-

zona, USA, 23-27 Jan. 2006.

[73] Australian Radiation Protection and Nuclear Safety Agency, “Radiation protection standard -

maximum exposure levels to radiofrequency fields - 3 kHz to 300 GHz,” Radiation Protection Series

No. 3, 2002.

[74] J. Rendon. (2005, 18 Jan.) Wal-Mart touts RFID results. [Online].

Available: http://searchmobilecomputing.techtarget.com/originalContent/0,289142,sid40

gci1048680,00.html?bucket=NEWS [12 Nov. 2007].

[75] D. W. Engels, and S. E. Sarma, “The reader collision problem,” in Systems, Man and Cybernetics,

2002 IEEE International Conference on, vol. 3, 2002, p. 6 pp. vol.3.

[76] B. Carbunar, M. K. Ramanathan, M. Koyuturk, C. Hoffmann, and A. Grama, “Redundant reader

elimination in RFID systems,” in Second Annual IEEE Communications Society Conference on Sensor

and Ad Hoc Communications and Networks (SECON), Santa Clara, California, USA, 2005.

[77] K. S. Leong, M. L. Ng, and P. H. Cole, “Dual-frequency antenna design for RFID application,”

in 21st International Technical Conference on Circuits/Systems, Computers and Communications (ITC-

CSCC 2006), Chiang Mai, Thailand, 10-13 July 2006.

[78] K. S. Leong, M. L. Ng, A. Grasso, and P. H. Cole, “Dense RFID reader deployment in Europe using

synchronization,” Journal of Communications, vol. 1, no. 7, pp. 9–16, 2006.

[79] K. S. Leong, M. L. Ng, and P. H. Cole, “RFID reader synchronisation,” in RFID Handbook: Applica-

tions, Technology, Security, and Privacy, S. Ahson and M. Ilyas, Eds. CRC, 2008.

[80] T. Skeie, S. Johannessen, and O. Holmeide, “Highly accurate time synchronization over switched

Ethernet,” in Emerging Technologies and Factory Automation, 2001. Proceedings. 2001 8th IEEE Inter-

national Conference on, 2001, pp. 195–204 vol.1.

Page 338

Bibliography

[81] CISC, “Consolidated proposal on synchronization,” ETSI Electromagnetic Compatibility and Radio

Spectrum Matteres, Task Group 34, RF Identification Devices, ETSI ERMTG34#12 012, 2006.

[82] Impinj, “RFID operation in Europe,” ETSI Electromagnetic Compatibility and Radio Spectrum Mat-

teres, Task Group 34, RF Identification Devices, ETSI ERMTG34#12 11r1, 2006.

[83] Intermec, “Smart Listen Before Talk,” ETSI Electromagnetic Compatibility and Radio Spectrum Mat-

teres, Task Group 34, RF Identification Devices, ETSI ERMTG34#12 010, 2006.

[84] RFID Journal. (2006, 10 Oct.) ETSI tests show EPC scaleable in Europe. [Online]. Available:

http://www.rfidjournal.com/article/articleview/2712/1/1/ [12 Nov. 2007].

[85] C. A. Balanis, Antenna theory: analysis and design, 3rd ed. John Wiley and Sons, 2005.

[86] R. C. Johnson, and H. Jasik, Antenna engineering handbook, 3rd ed. McGraw-Hill, 1993.

[87] C. A. Balanis, Antenna theory: analysis and design, 2nd ed. John Wiley and Sons, 1997.

[88] H. A. Wheeler, “The Radiansphere around a Small Antenna,” Proceedings of the IRE, vol. 47, no. 8,

pp. 1325–1331, 1959.

[89] K. Eshragain, “Electromagentic traffic sensing and surveillance,” Ph.D. dissertation, University of

Adelaide, 1980.

[90] I. of Electrical and E. Engineers, “Ieee standards definitions of terms for antennas,” Tech. Rep.,

1993.

[91] Cushcraft Corporation, “Antenna properties and their impact on wireless system performance,”

White Paper, 2002.

[92] J. Bernhard, and E. Michielssen, “Part B: antenna technology and numerical mehtods: antenna

parameters, various generic antenna and feed systems, and available software,” in Handbook of

Antennas in Wireless Communications, L. C. Godara, Ed. CRC Press LLC, 2002.

[93] R. J. Smith, Circuits, devices, and systems, 3rd ed. John Wiley & Sons, 1976.

[94] Texas Instruments. (2006, Oct.) UHF Gen2 STRAP. [Online]. Available: http://www.ti.com/rfid/

docs/manuals/pdfSpecs/RI-UHF-STRAP DataSheet.pdf [12 Nov. 2007].

[95] Texas Instruments, “HF antenna design notes,” Technical Application Report, 2003.

[96] P. V. Nikitin, and K. V. S. Rao, “Performance limitations of passive UHF RFID systems,” in Anten-

nas and Propagation Society International Symposium 2006, IEEE, 2006, pp. 1011–1014.

[97] M. L. Ng, K. S. Leong, and P. H. Cole, “Analysis of constraints in small UHF RFID tag design,”

in IEEE 2005 International Symposium on Microwave, Antenna, Propagation and EMC Technologies for

Wireless Communications, Beijing , China, 8-12 Aug. 2005.

Page 339

Bibliography

[98] D. C. Ranasinghe, M. L. Ng, K. S. Leong, and P. H. Cole, “Small UHF RFID label antenna design

and limitations,” in IEEE International Workshop on Antenna technology : Small Antennas and Novel

Metamaterials, New York, 6-8 Mar. 2006.

[99] S. R. Best, “A discussion on the quality factor of impedance matched electrically small wire an-

tennas,” Antennas and Propagation, IEEE Transactions on, vol. 53, no. 1, pp. 502–508, 2005.

[100] R. M. Fano, “Theoretical limitations on the broadband matching of arbitrary impedances,” Re-

search Laboratory of Electronics, Massachusetts Institute of Technology, Technical Report No. 41, 1948.

[101] Cushcraft Corporation, “Antenna polarization considerations in wireless communications sys-

tems,” White Paper, 2002.

[102] Texas Instruments. (2005, Nov.) Tag-it HF-I standard transponder IC - reference guide. [Online].

Available: http://www.ti.com/rfid/docs/manuals/refmanuals/HF-IStandardICRefGuide.pdf

[12 Nov. 2007].

[103] NXP, “SL2 ICS20 Smart Label IC Short Form Specification,” 2002.

[104] Ansoft Corporation, “User’s guide - High Frequency Structure Simulator,” 2003.

[105] T. H. Hubing, “Survey of numerical electromagnetic modeling techniques,” Technical Report for

Intel Corporation, TR91-1-001.3, 1991.

[106] Ansoft Corporation, “Introduction to scripting in HFSS,” 2003.

[107] V. C. Ramasami. (2007) HFSS MATLAB scripting API. [Online]. Available: http://www.ittc.ku.

edu/∼rvc/projects/hfssapi/doc/hfss-matlab-api.html [18 Jan. 2007].

[108] T-Tech Inc. (2007) PCB prototyping, autoSketch, quick circuit system, ISOPro software. [Online].

Available: http://www.t-tech.com/default.asp [12 Nov. 2007].

[109] Y. Lee, “Antenna circuit design for RFID applications,” Microchip Technology Inc. Application Note,

vol. AN710, 2003.

[110] Texas Instruments, “HF antenna cookbook,” Technical Application Report, 2004.

[111] S. S. Basat, K. Lim, J. Laskar, and M. M. Tentzeris, “Design and modeling of embedded 13.56 MHz

RFID antennas,” in Antennas and Propagation Society International Symposium, 2005 IEEE, vol. 4B,

2005, pp. 64–67 vol. 4B.

[112] S. C. Q. Chen, and V. Thomas, “Optimization of inductive RFID technology,” in Electronics and the

Environment, 2001. Proceedings of the 2001 IEEE International Symposium on, 2001, pp. 82–87.

[113] H. Greenhouse, “Design of planar rectangular microelectronic inductors,” Parts, Hybrids, and Pack-

aging, IEEE Transactions on, vol. 10, no. 2, pp. 101–109, 1974.

[114] Zork. (2007) The alternative to cork. [Online]. Available: http://www.zork.com.au/ [12 Nov.

2007].

Page 340

Bibliography

[115] ATMEL Corporation, “Datasheet - ATA5558 - 1kbit R/W IDIC with deterministic anticollision,”

2005.

[116] B. Mennecke, and A. Townsend. (2005, May) Radio frequency identification tagging as a

mechanism of creating a viable producer’s brand in the cattle industry. [Online]. Available:

http://www.card.iastate.edu/publications/DBS/PDFFiles/05mrp8.pdf [12 Nov. 2007].

[117] Allflex Australia. (2007) Allflex 36T NLIS tags. [Online]. Available: http://www.allflex.com.au/

nlis tags.htm [12 Nov. 2007].

[118] K. S. Leong, M. L. Ng, and P. H. Cole, “Investigation on the deployment of HF and UHF RFID

tag in livestock identification,” in IEEE Antennas and Propagation Society International Symposium,

Honolulu, Hawaii, USA, 10-15 Jun. 2007.

[119] K. S. Leong, M. L. Ng, and P. H. Cole, “HF and UHF RFID tag design for pig tagging,” in 11th

Biennial Conference of the Australasian Pig Science Association (APSA), Brisbane, Australia, 25-28 Nov.

2007.

[120] K. V. S. Rao, P. V. Nikitin, and S. F. Lam, “Antenna design for UHF RFID tags: a review and a

practical application,” Antennas and Propagation, IEEE Transactions on, vol. 53, no. 12, pp. 3870–

3876, 2005.

[121] M. L. Ng, K. S. Leong, D. M. Hall, and P. H. Cole, “A small passive UHF RFID tag for livestock

identification,” in IEEE 2005 International Symposium on Microwave, Antenna, Propagation and EMC

Technologies for Wireless Communications, Beijing, China, 8-12 Aug. 2005.

[122] Alien Technologies. (2007) 915MHz EPC Class 1 RFID tags. [Online]. Available: http:

//www.alientechnology.com/products/documents/alien 915mhz 128 bit.pdf [15 Jan. 2007].

[123] M. L. Ng, K. S. Leong, and P. H. Cole, “A small passive UHF RFID tag for metallic item identifica-

tion,” in 21st International Technical Conference on Circuits/Systems, Computers and Communications

(ITC-CSCC 2006), Chiang Mai, Thailand, 10-13 July 2006.

[124] K. S. Leong, M. L. Ng, and P. H. Cole, “Miniaturization of dual-frequency RFID antenna with

high frequency ratio,” in IEEE Antennas and Propagation Society International Symposium, Honolulu,

Hawaii, USA, 10-15 Jun. 2007.

[125] K. L. Wong, Compact and broadband microstrip antennas. John Wiley and Sons, 2002.

[126] W. Menzel, M. Al-Tikriti, and M. B. E. Lopez, “A common aperture, dual-frequency printed an-

tenna (900 MHz and 60 GHz),” Electronics Letters, vol. 37, no. 17, pp. 1059 – 1060, 2001.

[127] J. Collins. (2004, 1 Sep.) New two-frequency RFID system, RFID Journal. [Online]. Available:

http://www.rfidjournal.com/article/articleview/1105/1/1/ [12 Nov. 2007].

Page 341

[128] K. S. Leong, M. L. Ng, and P. H. Cole, “Investigation of RF cable effect on RFID tag antenna

impedance measurement,” in IEEE Antennas and Propagation Society International Symposium, Hon-

olulu, Hawaii, USA, 10-15 Jun. 2007.

[129] O. Staub, J.-F. Zrcher, and A. Skrivervik, “Some considerations on the correct measurement of

the gain and bandwidth of electrically small antennas,” Microwave and Optical Technology Letters,

vol. 17, no. 3, pp. 156–160, 1998.

[130] J. DeMarinis, “The antenna cable as a source of error in EMI measurements,” in Electromagnetic

Compatibility, 1988. Symposium Record. IEEE 1988 International Symposium on, 1988, pp. 9–14.

[131] C. Icheln, M. Popov, P. Vainikainen, and S. He, “Optimal reduction of the influence of RF feed

cables in small antenna measurements,” Microwave and Optical Technology Letters, vol. 25, no. 3,

pp. 194–196, 2000.

[132] C. Icheln, J. Krogerus, and P. Vainikainen, “Use of balun chokes in small-antenna radiation mea-

surements,” Instrumentation and Measurement, IEEE Transactions on, vol. 53, no. 2, pp. 498–506,

2004.

[133] Z. N. Chen, N. Yang, Y.-X. Guo, and M. Y. W. Chia, “An investigation into measurement of handset

antennas,” Instrumentation and Measurement, IEEE Transactions on, vol. 54, no. 3, pp. 1100–1110,

2005.

[134] S. Kahng, “A stack of metal rings for reducing common-mode current on a wire passing through

an aperture,” IEICE Trans Commun, vol. E88-B, no. 9, pp. 3819–3822, 2005.

[135] S. Saario, D. V. Thiel, J. W. Lu, and S. G. O’Keefe, “An assessment of cable radiation effects on

mobile communications antenna measurements,” in Antennas and Propagation Society International

Symposium, 1997. IEEE., 1997 Digest, vol. 1, 1997, pp. 550–553 vol.1.

[136] R. N. Simons, Coplanar waveguide circuits, components, and systems. John Wiley & Sons, 2001.

Page 342

Antenna Positioning Analysis and Dual-Frequency Antenna ...€¦ · Antenna Positioning Analysis...

Documents

Transcript of Antenna Positioning Analysis and Dual-Frequency Antenna ...€¦ · Antenna Positioning Analysis...