“Minimal Transmission Power vs. Signal Strength as ... · 4 4 Frank Reichenbach University of...
Transcript of “Minimal Transmission Power vs. Signal Strength as ... · 4 4 Frank Reichenbach University of...
“Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks”
International Workshop on Wireless Ad-hoc and Sensor Networks (IWWAN 2006), New York
Jan Blumenthal, Frank Reichenbach, Dirk Timmermann
Institute of Applied Microelectronics and Computer EngineeringUniversity of Rostock
June 30, 2006
22
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Outline
Problem StatementClassificationBackgroundDistance Estimation TechniquesNew Approach: Minimal Transmitting PowerWeighted Centroid Localization AlgorithmResultsConclusion
33
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Problem Statement
PreconditionsHundreds of sensor nodes are randomly deployedPosition initially unknown
Why do we need localization?Position ↔ MeasurementSelf organization, -healingGeographic Routing
ConsiderationsNodes miniaturizedNodes strongly resource limitedChanging dynamic topologyNodes are error-prone
44
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Conceivable Solution
Integrating existing localization system on every nodeGPS, GSM, Galileo
But:Sensor nodes are strongly resource limited
GPS has a relatively high power consumptionSensor nodes have to be tiny
GPS modules are comparatively largeLocalization availability
GPS does not work everywhereNodes must be cheap
GPS costs additionally
55
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Problem Solution
Equip only some nodes with e.g. GPSBeacons
Rest of the nodesUnknowns
Estimation of the position withDistancesAngles
αβ
γ
d1
d2d3
66
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Why not just using Trilateration?
Only 3 beacons needed (2D)Simple to calculateBut:
Distance estimations are highly defective!
RSSI over distance indoor (868MHz)
77
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Classification
Exact Localization(fine-grained)
Approximate Localization(coarse-grained)
• Angulation• Trilateration• Least Squares• Kalman Filter
Examples:• Dynamic Fine Grained Multilateration
(Savvides et al.)• Acoustic with Least Squares (Kwon et al.)
Examples:• Coarse Grained Localization (Bulusu)• APIT (He et al.)• Weighted Centroid Localization
(Blumenthal et al.)• Convex Position Estimation
(Doherty et al.)
• Geometric• Proximity• Scene analysis
88
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Observation Techniques
NeighboringNodes Distance Estimation Angle
Estimation
TimeMeasurement
SignalMeasurement
TimeMeasurement
„Time of Arrival“(ToA)
„Time Differenceof Arrival“(TDoA)
„Received Signal Strength Indicator“
(RSSI)
„Angle of Arrival“(AoA)
„Minimal Transmission
Power“
99
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Distance Estimation with RSSI in Theory
Received Signal Strength Indicator supported by hardware
Cheap and always availableCircuit measures the received energy of a signalCompared to a reference voltageReceived Power:
Logarithm:
20
4R
R SS
P G GP d
λπ
⎛ ⎞= ⎜ ⎟⎝ ⎠
20
4R
R SS
P G GP d
λπ
⎛ ⎞= ⎜ ⎟⎝ ⎠
[ ]2
0( ) [ ] 10 log 20 log( )4R S R SP d dBm P dBm G G dλπ
⎡ ⎤⎛ ⎞= + ⋅ − ⋅⎢ ⎥⎜ ⎟⎝ ⎠⎢ ⎥⎣ ⎦
1010
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Problems with RSSI in Practice
Transceiver sensitivitySignals in real world are strongly influencedAttenuation when passing objects
876MHz 8-20dB by a tree2.4 GHz bricks 3dB, tinted glass walls 19dB
Signal propagation characteristics can change frequentlyReceived Signal Strength depends on battery level
Blocking Reflection Scattering Diffraction
1111
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Result Measurements: Indoor (Chipcon)
Chipcon CC1010
1212
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Distance Estimation with Power Adjustment
Estimation of the minimal transmitting power of a beaconTransmitting power corresponds to the distanceTransmitting power P is a register SFR in the hardwareTransmitting power SFR adjustable in the interval SFR=0..100 (d=1-300m)
SFR=11
2m
SFR=14 SFR=16
Distance d to nodes
SFRmin
P
1313
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Result Measurements: Indoor (Scatterweb)M
in. S
endi
ng R
egis
ter V
alue
SFR
50 400100 150 200 250 300 350-5
0
5
10
15
20
25
30
35
Distance [cm]
1414
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Result Measurements: Linearized
What is d?
0
100
200
300
400
500
600
700
800
900
5 30 55 80 105 130 155 180 205 230 255 280 305 330 355 380Distance [cm]
Min
. Sen
ding
Reg
iste
r Val
ueSF
R
22 ( )( ) 1.3, 0
1.3SFR dSFR d m d n d with m n= ⋅ + → = = =
1515
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Weighted Centroid Localization (WCL)
Approach:Including distances in localizationDefine a function for the weight wij(d)
( )1
1
( , )''( , )
b
ij jj
i b
ijj
w B x yP x y
w
=
=
⎛ ⎞⋅⎜ ⎟
⎝ ⎠=⎛ ⎞⎜ ⎟⎝ ⎠
∑
∑
1
1'( , ) ( , )n
i jj
P x y B x yn =
= ∑
WCL
CGLCD
''( , )iP x y'( , )iP x y
1B 2B
3B4B
4id3id
2id1id
( , )iP x y
1616
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Determination of the Weight
Free parameter is the weightSubstituting the weight:
Optimal degree g?DensityPlacement
Simulations g = 3 is optimalReal g = 2
( )1
1
( , )''( , )
b
ij jj
i b
ijj
w B x yP x y
w
=
=
⎛ ⎞⋅⎜ ⎟
⎝ ⎠=⎛ ⎞⎜ ⎟⎝ ⎠
∑
∑
( )( )1
ij g
ij
wd SFR
=
1717
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Realization of a Localization
Beacons transmit position with increasing transmitting powerNode saves minimal transmitting powerBeacon reached max. transmitting power
counter for the round is incrementedRound based system via “Token Ring”-mechanism
Beacon (known position)Sensor node (unknown position)
B3
B1 B2
B4
1818
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Demonstrator ShowGraphical User Interface: SpyGlass (University of Luebeck)Transceiver Hardware: Scatterweb (FU Berlin)4 beacons at all corners, one unknown node was movedField size = 2x2 mYields in a running localization with 0,3m absolute error
http://rtl.e-technik.uni-rostock.de/~bj/movies/PositionEstimationUsingMinimalTransmissionPower.mpg
1919
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Discussion
ProsIncomplex Calculation (WCL) O(n)Fully distributed on every sensor node autonomouslyTransmission activity only on beaconsRelatively robust against defective inputs
ConsErrors arise by non-circular bordersRound-based transmission stresses the channelLatencyLimited precisionStrongly hardware dependent (Chipcon not possible)Transmission range depends on battery level
2020
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Conclusions and Future Work
Localization in WSN is strongly demandedMost classical observation techniques are defective (ecsp. indoor)
New Approach: Minimal transmitting power
Verification:High resolution and smaller variances than RSSICombining localization algorithm WCL with MTP showed good practical results2x2m field with 9 areas max. localization error = 0.3m
2121
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Thank You!
2222
Frank ReichenbachUniversity of Rostock
„Minimal Transmission Power vs. Signal Strengthas Distance Estimation for Localization in
Wireless Sensor Networks“
IWWAN 2006
Substituting the Parameters
v
2
( )4E SP P S
rλπ
⎛ ⎞= ⎜ ⎟⎝ ⎠
Sr P=d
Ps2
( )4E SP P S
rλπ
⎛ ⎞= ⎜ ⎟⎝ ⎠
2
( )4E SP P S
dλπ
⎛ ⎞= ⎜ ⎟⎝ ⎠
2( )Sd P S=
? SS P=