APPL: Anchor Path Planning – based Localization for Wireless Sensor Networks

APPL: Anchor Path Planning –based Localization for Wireless Sensor Networks

Imane BENKHELIFA and Samira MOUSSAOUI

LSI, Computer Science DepartmentHouari Bourmediene University of Science and Technology

Algiers, Algeria

222/07/2011Imane BENKHELIFA

Mosharaka International Conference on Communications, Computers and Applications (MIC-CCA2011)

Motivation & IntroductionLocalization in Wireless Sensor NetworksLocalization using a Mobile AnchorAPPL: Anchor Path Planning based LocalizationEvaluationConclusion & Future Directions

Outline

3

Motivation & IntroductionLocalization in Wireless Sensor NetworksLocalization using a Mobile AnchorAPPL : Anchor Path Planning based LocalizationEvaluationConclusion & Future Directions

MotivationIntroduction

22/07/2011

Battlefield Surveillance

Maritime SurveillanceForest Fire Detection

Precision Agriculture Monitoring Patients

Monitoring Animals

Imane BENKHELIFA Mosharaka International Conference on Communications, Computers and Applications

(MIC-CCA2011)

4


MotivationIntroduction



• WSN Attractive Caracteristics (Fast Deployment, Fault Tolerance, reduced cost,… etc)

• Why Sensor Localization is important?

A detected event is only useful if an information relative to its geographical position is provided




Outline

6


IntroductionRSSI TechnologyAoA Technology



• Localization = to assign accurate or estimated position to a sensor

• Why not a GPS (Global Posionning System) in each node?

• Solution: to equip some nodes with GPS Anchors

• Measurement Technologies:ToA (Time of Arrival)

TDoA (Time difference of Arrival)

RSSI (Received Signal Strength Indicator)

AoA (Angle of Arrival)

Costly

7





• RSSI (Radio Signal Strenght Indicator)

• Considers the power loss of a signal between its emission and reception

• This loss varies with the distance between two sensors. More the sensors are distant (resp. close), more the loss is important (resp.weak)

• This loss is then converted to a distance

8





• AoA (Angle of Arrival)

Each sensor has an oriented antenna and a reference axis

Absolute AoA :(θi +Δθ )(mod 2 π), i = {1,2}




Outline

10

Motivation & IntroductionLocalization in Wireless Sensor NetworsLocalization using a Mobile AnchorAPPL : Anchor Path Planning based LocalizationEvaluationConclusion & Future Directions

IntroductionScan Trajectory

Double Scan TrajectoryHilbert Trajectory



• Many anchors equipped with GPSEnergy comsuption & high costs

• Solution: using a mobile anchor (robot, plane, drone, vehicle, tractor…)

11






• Other than serving as a mobile anchor, the mobile node can perform other tasks necessary for the network operations

Tasks:• Reconfigures or recalibrates sensors, • Synchronizes the clocks, • Collects data from the sensors ,• Deploys new sensors, and disables existing ones

12






Method using a mobile Anchor [Kouts. 07]• Goal:

Planning the anchor trajectory Reducing the Localization error and Time

• Principle: Based on RSSI (conversion of signal to distance) Bayesian Inference (Probabilistic approch)

• Trajectories:- SCAN- DOUBLE SCAN- HILBERT

13






D= (L/R+1)x L+(L/R)x R= (L/R+2)x L

Advantages:

Good coverage of nodes Short traveled distance Reduced localization time

Drawbacks:

Abscissa-error is bigger than ordinate- error if resolution >> transmission range

Straight lines introduce collinearity

L

R

14






D= 2x DSCAN = 2 (L/R+2)x L

Advantages:

Scans 2 dimensions (horizontal & vertical) Ensures error-similarity along both axis

Drawbacks:

Double distance / resolution Higher localization time/errorL

R

15






D= (L/R+2)x L +R

Advantages:

Traveled distance differs only in R from SCAN Diffusion of non-colinear messages Similar error between x/y axis

Drawbacks:

Some nodes cannot receive anchor messages (square holes)

L

R




Outline

17

Motivation & IntroductionLocalization in Wireless Sensor NetworsLocalization using a Mobile AnchorAPPL: Anchor Path Planning based LocalizationEvaluationConclusion & Future Directions

Estimation TechniqueAnchor Path Planning

Squares TrajectoryArchimedian Spiral Trajectory

Waves Trajectory



• APPL (Anchor Path Planning based Localization)Principle:– One single Mobile Anchor– Anchor Trajectory is chosen in advance– At Each reception of anchor message, the static node rebuilds

an Anchor Box– The Anchor Box is built based on distances deduced from the

received RSSI– The node draws N samples from this box– The estimated position is the mean of samples

18




Waves Trajectory



• Building the Anchor Box

Actual Position

Estimated Position

r1

r2

r3

19




Waves Trajectory



• Anchor Path PlanningTo offer network coverage any deterministic trajectory with

properly selected parameters

To provide good quality beacons non-collinear messages, and signal strengths can be mapped to their distances from the sensor

•Anchor Trajectories- Squares- Archimedean Spiral- Waves

20




Waves Trajectory



• Squares Trajectory

20

Advantages:

Same traveled distance as ScanGood node coverage2 dimentional scanSimilar error between x/y axis

Drawbacks: Straight lines introduce collinearity if

resolution >> transmission range

L

R

L

D= R+4 [(L-2R)+(L-4R)+(L-6R)+….+(L-2[L/2R]R)] = (L/R+2)x L

Imane BENKHELIFA Mosharaka International Conference on Communications, Computers and Applications (MIC-CCA2011)

21




Waves Trajectory

21

• Archimedean Spiral Trajectory

21

Advantages:

Traveled distance less than SCAN with + 37% Localization time reduced Reception of non-colinear messages Similar error between the two axis

Drawbacks:

Coverage of the four corners

If the deployment area increases, turns become larger the amount of non- collinearity reduces

R

Distance ≤ Rπn/4 (4n-1) with n= [L/2R] + 122/07/2011

Imane BENKHELIFA Mosharaka International Conference on Communications, Computers and Applications (MIC-CCA2011)

22




Waves Trajectory

22

• Waves Trajectory

22

3R/2

R

Advantages:

Reception of many non-colinear messages from only one segment Traveled Distance less than SCAN with 16% Good network coverage x/y error similarity

D=2R/3 ( [ 2L/3]-1 ) + (1+ ([L/R ]-1) ([ 2L/3R ])) (π R/2 )22/07/2011




Outline

24


Simulation EnvironmentEstimation Technique Evaluation

Path Length and Travel TimeLocalization Coverage

Localization Error

2424

• Simulation Environment:– NS2 Simulator on Linux (Ubuntu 9.10)– Deplyment area =200m x 200m– Number of nodes =100– Transmission range =30m– Anchor speed =20m/s

• Evaluation Metrics:– Path Length and Travel Time– Localization Coverage (Number of localized nodes)– Localization Error (Distance between estimated position and real

position)22/07/2011

Imane BENKHELIFA Mosharaka International Conference on Communications, Computers and Applications

(MIC-CCA2011)

25




Localization Error



Path =Squares, Resolution =20m

10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

70

80

90

100

Building Boxes

RSSITransmission range

% n

odes

with

err

or <

= 0.

1 r

(% end time)

26




Localization Error



Path =Squares, Resolution =20m

10 20 30 40 50 60 70 80 90 1000

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

Diffusion Interval

1s 3s

mea

n er

ror

(r)

(% end time)

27




Localization Error



Resolution =20m

Trajectoriy Path Length (m) Travel Time (s) Gain over Scan

Double Scan 4800 240 -100%

Hilbert 2420 121 -0,82%

Scan 2400 120 0%

Squares 2400 120 0%

Waves 2003,6 100,18 +16,5%

Spiral 1500 75 +37,5%

28




Localization Error



Resolution =60m

10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

70

80

90

100

SquaresScanHilbertSpiralWavesArbitrary

(% end time)

Estim

ated

nod

es (%

)

29




Localization Error



Path= Squares

10 20 30 40 50 60 70 80 90 1000

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

Resolution variation

R=20m R=40m R=60m

(% End time)

mea

n er

ror (

r)

30




Localization Error



Resolution =60m

Waves Squares Spiral Scan Hilbert0

0.05

0.1

0.15

0.2

0.25

0.3m

ean

erro

r (r)

31




Localization Error



• x/y error gap

Resolution =60m

Waves Squares Spiral Scan Hilbert0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

x-errory-errormean error

(r)




Outline

33


ConclusionFuture Directions



• APPL uses a very simple localization algorithm that makes a good combination of RSSI measurements and anchor positions

• Proposed Trajectories (SQUARES, ARCHILEDEAN SPIRAL and WAVES) are specifically designed with better localization accuracy, coverage and path length in mind

• Squares gives the quickest coverage• Sipral gives the shortest distance• Waves gives the best accuracy

• When the path resolution is much larger than the transmission range, the collinearity becomes more significant. In such cases, WAVES outperforms previously proposed path types.

34


ConclusionFuture Directions



• The static paths work well when the sensors are assumed to be uniformly deployed. However, in the cases where such assumption is not valid, there is a strong intensive to dynamically adjust the path during the localization procedure

Thank you

APPL: Anchor Path Planning – based Localization for Wireless Sensor Networks

Documents

sensor localization

anchor path

wireless sensor networks

mobile anchorappl

localization methods

localization basics

mobile anchorafter

fault tolerance