Positioning in Ad-Hoc Networks-
A Channel Model for Low Power Data Transmission
Jan BeutelComputer Engineering and Networks Lab
Swiss Federal Institute of Technology (ETH) Zurich
June 15, 2001
Computer EngineeringComputer Engineeringand Networks Laboratoryand Networks Laboratory
2
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Outline
• Sensor Networks
• Radio Channel• Propagation Channel
• Simulation using ETH Raytracing Tool
• Some RSSI Sample Data• First Positioning Results
3
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Sensor Networks
• Ad-Hoc network of many sensors, monitors and actuators (>100 total)
• Vision of system on a chip• No infrastructure• Deployable in different
environments• Limited radio range• Node clusters and depletion
Goal: Ultra Low Power
4
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Radio Ranging Methods
TOA – time of arrival
Carrier phase
AOA – angle of arrival
Signal strength
TDOA – time distance of arrival
5
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Radio Frontend Requirements
• Simple direct down-conversion frontend
• Emphasis on digital processing (µJ vs. pJ per Bit)
• Mostly sleepy radio (receiver)• Tx power scalable (Wolisz et. al.)
• CMOS SOC’s -> 2.4 GHz / 5 GHz• PTx < 0 dBm
ReconfigurableDataPath
ReconfigurableState Machines Embedded uP
FPGA
DedicatedDSP
• RSSI used in transceivers for carrier sensing -> available at no cost
• Possibly separation of multi-path and multi-access through advanced baseband processing (RAKE)
6
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Propagation Channel in Mixed Environments
• Mechanism: Reflection, Diffraction, Scattering
• Transmission through a wall:– up to 6-10 dB at 2.5 GHz -> walls are not transparent
• Received signal strength for low power at 1/r3 to 1/r4. In commercial space, e.g., supermarkets, Lucent measured 1/r3.8 at 2.5 GHz. (Source G. Wright, Lucent)
• High impact of geometry on overall picture• Local impact of material fluctuates statistically
• Models are highly specific or give average overview
7
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Received Signal Strenght
• Multi-path vs. LOS components• Delay and AOA reflect geometry (large scale)• Fading• Amplitude influenced by dielectrics, angle of
incidence and polarization
• Multiple access interference, depending on CD/TDMA scheme
• Around λ/2 spacing, signal strength variation may be as high as 30-40 dB (small scale fading)
8
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Tools for Simulation: Raytracing
• Exact modeling of LOS path and reflections• Differentiating
– Order of reflections– Geometry– Material– Polarization– Delay and AOA– Amplitudes– Power Spectrum Densities
• Raytracing tool developed by J. Hansen
9
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Power Density: Polarization
• Transmitter located at ceiling• Arbitrary location receiver
Distance
Elevation
Distance
Elevation
10
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Power Density: Windows
• Small window (0.7 x 0.8) in wall• Windows behave like walls
Distance
Elevation
Distance
Elevation
11
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Power Density: Room Size and Layout
• Transmitter fixed in middle of Z axis• Guidance effect of very long room
Distance
Elevation
Distance
Elevation
12
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Bluetooth RSSI Samples Free Space
-25
-20
-15
-10
-5
0
5
10
15
201 30 59 88 117
146
175
204
233
262
291
320
349
378
407
436
465
494
523
552
581
610
639
668
697
726
755
784
813
842
871
900
929
958
987
1016
1045
Samples/Distance
RSS
I
Series1Series2
S
13
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Bluetooth RSSI Samples Office Space
-25
-20
-15
-10
-5
0
5
10
15
201 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 103
109
115
121
127
133
139
145
151
157
163
169
175
181
187
193
199
Samples/Distance
RSS
I
Series1Series2Series3
14
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Wavelan 802.11b RSSI Samples
-120
-100
-80
-60
-40
-20
0
20
40
60
80
-3 -2 -1 0 0 1 2 3 4 5 6 6 7 8 9 10 11 12 12 13 14 15 16 17 18 18 19 20 21 22
Distance [m]
Sign
al a
nd N
oise
Lev
el [d
Bm
]
LSNR Avg LSL Avg LNL Avg RSNR Avg RSL Avg RNL Avg Path Loss 1/r 2̂Path Loss 1/r 4̂
15
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Wavelan 802.11b Indoor Environment
16
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Redundant Triangulation
Iteration on 25 individual ranges with 50% each
0 0.5 10
0.2
0.4
0.6
0.8
1Delaunay Mesh of 25 Networked Nodes
x0 0.5 1
0
0.2
0.4
0.6
0.8
1Solution on 25 Ranges and 50% Error
x
0 0.5 10
0.2
0.4
0.6
0.8
150 Solutions and Mean
x0.4 0.45 0.5 0.55 0.6
0.4
0.45
0.5
0.55
0.6Zoom on Error
x
dx 0.0054dy 0.0058
1% error
17
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Triangulation on Real Data
• Application to Wavelan RSSI data shows good average
Solution• Iterative triangulation• Overload systems• Influence of
geometric DOP• Include environmental
information
Yielding sub meter DOP
18
ETH
Zur
ich
Jan Beutel, June 15, 2001Jan Beutel, June 15, 2001
Literature• Jan Hansen. A novel stochastic millimeter wave indoor radio
channel model. Journal on Selected Areas in Communication, 2001.• Chris Savarese, Jan Beutel, and Rabaey Jan M. Locationing in
distributed ad-hoc sensor networks. In Proceedings of ICASSP, 2001.
• Jan M. Rabaey, Josie Ammer, Julio L. da Silva Jr., Danny Patel, and Shad Roundy. PicoRadio Supports Ad Hoc Ultra-Low Power Wireless Networking. IEEE Computer, 33(7):42-48, July 2000.
• Theodore S. Rappaport. Wireless Communications. Prentice Hall, 1996.
• Danny Patel. Energy in Ad-Hoc Networking for the PicoRadio. Master's thesis, University of California at Berkeley, 2000.
• Jean Pierre Ebert and Adam Wolisz. Combined tuning of rf power and medium access control for wlans. In Proceedings of 1999 IEEE International Workshop on Mobile Multimedia Communications, pages 47-82, November 1999.
• Kamilo Feher. Wireless Digital Communications: Modulation and Spread Spectrum Applications. Prentice Hall, 1995.