MIT_Auto_ID_rv_2.1_Levesque.ppt

January 23, 2006 ©KP

RFID Academic Convocation AUTO-ID Laboratory at MIT RFID Academic Convocation AUTO-ID Laboratory at MIT

CWINS

Trends in RF Location Sensing Prepared by: Kaveh Pahlavan,

Prof. of ECE and CS and Director CWINS, WPI Presented by: Allan H. Levesque

January 23, 2006 RFID Research Convocation

Outline

Evolution of RF Location Sensing Applications

– in Commerce – in Military and Public Safety – in other disciplines

Localization Research at CWINS/WPI – RSS Localization– Accurate Localization Using TOA


Evolution of Localization Industry Problem of locating soldiers carrying mobile radios was first

addressed in World War II During the Vietnam conflict US department of defense lunched a

series of GPS satellites to support military operation In 1990 signals from GPS satellites became accessible to public In 1996 FCC required locating mobile phones with in 100 meters in

67% of locations In 1997 urban and indoor geolocation started

– Military: SUO/SAS project– Commercial: PinPoint, WearNet

In 2000 indoor WiFi localization (Newbury Networks and Ekahau) In 2003 UWB localization for WPANs In 2004 localization for sensor networks In 2005 outdoor WiFi localization as a GPS alternatives

– Commercial: WiFi positioning in outdoor areas (Skyhook)– Military: Signals of opportunity for disaster recovery (DARPA)


Localization in Commerce


Typical Commercial Applications

Asset tracking in warehouses and hospitals Dispatching and monitoring support staff Locating elderly persons Locating children in public areas Emergency management in hospitals Invisible fencing of pets Tracking golf balls Locating surgical equipment in operation rooms Guiding visitors in museums Navigating sight impaired persons GPS complement for indoor applications


PC

Laptop

Printer

Fax

Keyboard

Mouse

Scanner

PDA CRT projector

Video

Pen computer

Hand held computer

Cell phone

iBook

Today Everything is a Terminal!

PROGRAM

MonitorLocation Sensors

PBX

Appliances/Security

Computer/Communication

VCVCRR

Entertainment /Home Networking

Control/Metering


Standards and Terminals

1

100,000

10

100

1,000

10,000

MP

EG

2 V

ideo

3D G

ames

Ba

nd

wid

th (

Kb

ps

)

(100 Mbps)

(10 Mbps)

(1 Mbps)

Pri

nti

ng

Inte

rnet

Acc

ess

Sec

uri

ty S

yste

ms

Tra

dit

ion

al G

ames

Ph

on

e A

pp

lian

ces

Rea

l A

ud

io

MP

3 A

ud

io

Sm

art

Ap

pli

ance

s

Uti

lity

Met

erin

g

Cam

cord

er D

V

US

B C

on

nec

tors

1000,000 (1 Gbps)

Mu

ltip

le H

DT

V

802.15.4 WPAN

802.11 WLAN

802.15.3a WPAN


Mbps1 10 1000.1

Out

door

Fixed

Walk

Vehicle

Indo

or

Fixed/Desktop

Walk

Mobility

WLAN

User Bitrate, Datacom services

WPAN

E-911 (100 meters 67%)

WiFi Positioning (a few meters indoors)

UWB (a few centimeters indoor)

Localization Accuracy

1000

3G Cellular

2G C

ellu

lar


Accurate Localization in Military and Public Safety


Typical Applications

Navigation for fireman Navigation for policemen Providing situation awareness for

soldiers Tracking inmates in prisons Tracking in disaster recovery zones


SUO-SAS and Urban Fighting


Robot in rubble

Localization Everywhere

Forces inside buildings

Unattended ground sensorsUrban canyons

Inside caves


Using Signals of Opportunity


Localization in other Research Areas

Location-based handoff Location-based routing in ad hoc networks Location-based authentication and security


Localization Research at CWINS/WPI


Two Technologies RSS

– Simple hardware • Accuracy is not sensitive to multipath and bandwidth• Does not need synchronization• Can work with any existing infrastructure and card which can measure power

– For accurate localization • Needs high density of references or long training procedure• Needs more complex and computationally intensive algorithms

– Used for WiFi Positioning by Ekahau, Newbury Networks, PanGo, Skyhook. TOA

– Suitable for accurate indoor positioning• Only needs a few references• Does not need training

– More complex hardware • Accuracy is sensitive to multipath and bandwidth• Needs synchronization of the reference time

– Used in GPS, PinPoint, WearNet, 802.15.3 and 4.


Two Classes of Algorithms

Tx1Tx2

Tx

3

d1

d2

d3

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x x

x x

x x

x x

x x

x x

x x

x x

x x

x

AP1 AP2

AP3

AP4 AP5

MH

100 m

ReferencePoint

100

m

GridResolution

Y

X

Distance Based Localization

Pattern Recognition


RSS Localization at CWINS Microsoft: Nearest Neighbor algorithm based on signature data collection in a

training period.– P. Bahl and V. Padmanabhan, “RADAR: an in-building RF-based user location and

tracking system,” IEEE INFOCOM, Israel, March 2000. Ekahau: Uses a statistical algorithms for localization with improved accuracy

and reduced training data.– T. Roos, P. Myllymaki, H. Tirri, P. Miskangas, and J. Sievanen, “A Probabilistic

Approach to WLAN User Location Estimation,” International Journal of Wireless Information Networks, Vol. 9, No. 3, July 2002.

CWINS: Develops a real time laboratory testbed and a framework for comparactive performance evaluation of localization algorithms.

– M. Heidari and K. Pahlavan, Performance Evaluation of Indoor Geolocation Systems Using PROPSim Hardware and Ray Tracing Software, IWWAN’04, Oulu, Finland, May 2004.

CWINS: Uses channel modeling techniques (Ray Tracing and WiFi channel models) to eliminate the need for training data and improve the performance.

– A. Hatami, and Kaveh Pahlavan, " Comparative statistical analysis of indoor positioning using empirical data and indoor radio channel models," Consumer communications and networking conference, 2006.


Challenges for TOA Estimation

Coverage– Existing knowledge only supports coverage of the total power while in

TOA estimation we are keen on the coverage of the DP. We need more measurement and modeling effort to solve this problem.

Bandwidth– Common belief: increase in bandwidth increases the resolution

therefore UWB is the final solution– Problem: Increase in bandwidth reduces the coverage of the first

path, which means UWB is not the ultimate solution for Indoor Geolocation

UDP– Errors caused by UDP conditions are very large – Detection of UDP is difficult, so we can not avoid UDPs– If we detect the UDP and avoid, coverage become very small

Therefore: we need research in algorithms which operate in UDP conditions and provide good coverage with reasonable bandwidth


TOA Research at CWINS

Modeling of the behavior of TOA sensors (for NSF) UWB

– For high-resolution short distance (for DARPA)

Super-resolution algorithms– For signals of opportunity (with NSF and RCI)

Diversity Techniques– AOA (with Draper Laboratory)– Space and cooperative location estimation (with IWT)


Conclusion Localization is one of the fastest growing

industries attracting strong research investments from military and commerce

Variety of applications are emerging around RSS and TOA based technologies

CWINS is the oldest laboratory working in both RSS and TOA technologies with close ties to commerce and military

Localization for RFID applications is one the areas of interest at CWINS for future research

MIT_Auto_ID_rv_2.1_Levesque.ppt

Business

Transcript of MIT_Auto_ID_rv_2.1_Levesque.ppt