Positioning in Ad-Hoc Networks-

Directions and Results

Jan BeutelComputer Engineering and Networks Lab

Swiss Federal Institute of Technology Zurich

August 10, 2002

Computer EngineeringComputer Engineeringand Networks Laboratoryand Networks Laboratory

2

ETH

Zuri

ch

Jan Beutel, August 10, 2002Jan Beutel, August 10, 2002

Ad-Hoc Network Scenarios

•Low power•Small size•Very large population

•No infrastructure necessary•Varying population density•Multihop environment•Partitioning

3

ETH

Zuri

ch

Jan Beutel, August 10, 2002Jan Beutel, August 10, 2002

Positioning: The Problem

Finding the position of networking nodes

Relative vs. Absolute Positioning Mode

Reference Positions, Map

Database

Other Networking Nodes, Distance and Geometric

Topology

4

ETH

Zuri

ch

Jan Beutel, August 10, 2002Jan Beutel, August 10, 2002

-25

-20

-15

-10

-5

0

5

10

15

20

1 42 83 124 165 206 247 288 329 370 411 452 493 534 575 616 657 698 739 780 821 862 903 944 985 1026 1067

Series1

Series2

S

RSSI Samples Over Distance - Free Space

0

10

20

30

40

50

60

70

80

-3 -2 -2 -1 0 0 1 2 2 3 3 4 5 5 6 7 7 8 9 9 10 11 11 12 12 13 14 14 15 16 16 17 18 18 19 20 20 21 21 22

LSNR Avg

RSNR Avg 802.11b

Bluetooth

5

ETH

Zuri

ch

Jan Beutel, August 10, 2002Jan Beutel, August 10, 2002

Redundant Triangulation

Every node executes

•Identification of neighbors

•Establishing range estimates

•Maintaining a set of a minimum of 3 linear equations to the neighbors

•Solve for MMSE

Dissemination of data over the network

6

ETH

Zuri

ch

Jan Beutel, August 10, 2002Jan Beutel, August 10, 2002

Redundant Triangulation and Filtering

Average over 25 individual triangulations with 50% range error

0 0.5 10

0.2

0.4

0.6

0.8

1

Delaunay Mesh of 25 Networked Nodes

x

0 0.5 10

0.2

0.4

0.6

0.8

1

Solution on 25 Ranges and 50% Error

x

0 0.5 10

0.2

0.4

0.6

0.8

1

50 Solutions and Mean

x

0.4 0.45 0.5 0.55 0.60.4

0.45

0.5

0.55

0.6

Zoom on Error

x

dx 0.0054

dy 0.0058

1% position error

7

ETH

Zuri

ch

Jan Beutel, August 10, 2002Jan Beutel, August 10, 2002

Influence of Range Quantization

8

ETH

Zuri

ch

Jan Beutel, August 10, 2002Jan Beutel, August 10, 2002

Very Large Errors and Topology

3 anchors ~ 94%

4 anchors ~ 6%

5 anchors >1%

9

ETH

Zuri

ch

Jan Beutel, August 10, 2002Jan Beutel, August 10, 2002

Influence of Border Regions

Center I

Edge II

Corner III

10

ETH

Zuri

ch

Jan Beutel, August 10, 2002Jan Beutel, August 10, 2002

Influence of Border Regions

11

ETH

Zuri

ch

Jan Beutel, August 10, 2002Jan Beutel, August 10, 2002

Ad-hoc Network Simulation Environment

12

ETH

Zuri

ch

Jan Beutel, August 10, 2002Jan Beutel, August 10, 2002

The TERRAIN Algorithm .

• Triangulation via Extended Range and Redundant Association of Intermediate Nodes

• Algorithm creates local maps

• Every node waits to beincluded in 3 maps

• Extended rangescalculated fromrespective maps

• Triangulation node basedon extended ranges

• Network-wide iterations

1

2

3

radio range

extended range

intermediate node