Localization Techniques in Wireless Sensor Networks

1
Localization Techniques in Wireless Sensor
Networks
(z,t)
d3
d2
(u,v)
(x,y)
d1

• Prepared by Abdelmounaim Dahbi
• In partial fulfillment of the requirements for
  the course
• Wireless Ad Hoc Networking
• Instructor Professor Ivan Stojmenovic
• University of Ottawa

2
Outline

• Introduction
• Applications
• Beacon Nodes (Anchor Nodes)
• Distance/Angle Measurement Techniques
• Centralized Algorithms
• Distributed Algorithms
• Range-based Localization Techniques
• Range-free Localization Techniques
• Iterative Refinement
• Concluding Remarks
• Ongoing Research Issues
• References

3
Introduction

• What is Sensor Localization?
• The determination of the absolute or relative
  position of sensor nodes (geographical locations
  of sensors)

N3(x3,y3)
d3
d2
a
N2(x2,y2)
ß
y1
d1
N1(x1,y1)
x1
4
Introduction

• Why Sensor Localization?

• sensing data (Phenomena) without knowing the
  sensor location is most of the times meaningless

Large number of randomly scattered sensor nodes
5
Introduction

• Why Sensor Localization?

Gateway
This is true in any location sensitive
application, location aware service
6
Introduction

• Isnt GPS just the answer?
• Yes, but
• Not available indoor
• Limited in certain environments such as Bush
• Not accessible from under water
• Constraints on the cost of sensors
• Constraints on the size of sensors
• Constraints on the energy consumption
• Not very accurate

7
Applications

• Network Functions Geographical Routing,
  Collaborative Signal Processing
• Bush Fire Surveillance/Detection
• Intrusion Detection
• Habitat Monitoring/Wildlife Tracking (ZebraNet)
• Water Quality Monitoring
• Pollution Monitoring
• Traffic Monitoring
• Target Tracking (Military tracking enemy
  vehicles, and Civilian tracking wild animals in
  wildlife preserves)

8
Beacon Nodes (Anchor Nodes)

• Ordinary sensor nodes that know their global
  coordinates a priori
• Either hard-coded coordinates
• Or GPS equipped sensor nodes
• Different uses of beacon nodes (Reference,
  Flooding of their positions and other data)
• Importance of Beacon placement
• For 2D three and 3D four beacon nodes are needed
• But, costly

9
Distance/Angle Measurement TechniquesAngle of
Arrival (AoA)

• The angle between the propagation direction of an
  incident wave and some reference direction
• Does not require synchronization
• But, costly and requires extensive signal
  processing

10
Distance/Angle Measurement TechniquesReceived
Signal Strength Indicator (RSSI)

• In theory, the energy of a radio signal
  diminishes with the square of the distance from
  the signals source.
• Low cost all sensors have radios
• But in practice, RSSI ranging measurements
  contain noise (in the order of meters)
• Difference in propagation in different
  environments

11
Distance/Angle Measurement TechniquesTime of
Arrival (ToA)

• c The propagation speed of the radio signal
  (speed of light)
• Accurate
• But, requires precise synchronization

12
Distance/Angle Measurement TechniquesTime
Difference of Arrival (TDoA)

• c The propagation speed of the radio signal
• ss The propagation speed of the
  ultrasound/acoustic signal
• Accurate, No synchronization required
• But, costly (Hardware)

13
Distance/Angle Measurement TechniquesRadio Hop
Count (DV-Hop)

• Connectivity data
• hij Shortest path i,j (number of hops)
• dij Distance i,j
• dij lt R x hij
• Better estimate dhop
• dij hij x dhop

dhop avg hop distance 4
hAB3 hops
B
A
14
Distance/Angle Measurement TechniquesRadio Hop
Count (DV-Hop)

• nlocal The expected number of neighbors per node
• hij Length of the shortest path between sensor i
  and sensor j in terms of number of number of hops
• dij The Distance between sensor i and sensor j
• dhop The Average hop distance
• dij hij x dhop

15
Distance/Angle Measurement TechniquesRadio Hop
Count (DV-Hop)

• Distance measurements are always integral
  multiples of dhop
• Environmental obstacles can prevent edges from
  appearing in the connectivity graph that
  otherwise would be present
• Depends on the density nlocal

16
Centralized Algorithms

• Migration of internode ranging and connectivity
  data to a sufficiently powerful central base
  station
• Complex processing of the collected data
• Migration of the resulting locations back to the
  respective nodes.
• Examples
• SDP The SemiDefinite Programming
• MDSMAP MultiDimensional Scaling

17
Distributed Algorithms

• Each sensor collects its data
• Computation is done by the sensors
• Several iterations might be required
• Not as accurate as centralized but does no
  migrations between a central station and the
  sensors
• Examples
• Triangulation
• Trilateration/Multilateration
• Bounding Box (Min-Max)
• Centroid

18
Range-based Localization Techniques Triangulation
y

• AoA to compute the angles
• The number of BSs needed for the location process
  is less
• Compute the linear least-square solution
• Assuming (x1,y1)(0,0) and the x axis defined by
  the two beacon nodes we have

x
19
Range-based Localization Techniques
Trilateration/Multilateration

• Distance measured RSSI, ToA, TDoA
• Requires at least 3 BNs in 2D, and 4 BNs in 3D..
• Compute the linear least-squares solution
• Multilateration if more than three beacons are
  used to estimate the sensors position

20
Range-free Localization Techniques Bounding Box
(Min-Max)

• Distance based on Radio Hop Count (DV-Hop)
• Simple
• But less accurate

21
Range-free Localization Techniques Centroid
Algorithm
.
.

• Nodes localize themselves to the centroid of
  their proximate reference points

.
Xik,Yik
Xi,Yi
.
x
Xi1,Yi1
.
Xi2,Yi2
22
Iterative Refinement

• Node obtains initial position
• Node broadcasts its position
• Position is refined iteratively using
• Distances to neighbors
• Nodes previous positions

23
Concluding RemarksWhat is the best localization
algorithm?

• No best algorithm
• Depends on
• Error in range measurement
• Connectivity
• Network topology
• Node capabilities
• Application requirements
• ...

24
Concluding RemarksPros and Cons

• Two main types of distributed localization
  algorithms
• Range-based
• Estimating the coordinates based on the collected
  information of distances or angles among nodes
• Merit Relatively high accuracy
• Drawback Costly (Hardware, Power consumption)
• Range-free
• Estimating the coordinates based on the
  connectivity relations
• Merit Cost-effective
• Drawback Not as accurate (But coarse accuracy
  is sufficient for most sensor network
  applications)

Hardware/Energy Cost vs Location Precision
25
Ongoing Research Issues

• Noisy distance measurement
• Costly distance measurement (hardware, energy)
• Few beacons
• Scale
• Mobility

26
Questions

• Q1 Triangulation is based on the law of sines
  which states
• (sin a)/A(sin b)/B(sin c)/C
• Prove the law of sines
• Answer
• Sin a L/B, sin b L/A
• B . sin a A . sin b
• (sin a)/A(sin b)/B

27
Questions

• Q2 The Radio Hop Count (DV-Hop) distance
  estimation technique is based on the average hop
  distance dhop and the hop count hij (the length
  of the shortest path in the graph between si and
  sj in terms of the number of hops). This
  technique has a major drawback related to
  environmental obstacles which can prevent edges
  from appearing in the connectivity graph that
  otherwise would be present. Give an example of a
  graph where such drawback is highlighted..

Answer In this diagram, hAC 4. Unfortunately,
hBD is also four, due to an obstruction in the
topology.
28
Questions

• Q3 Knowing that dhop3 and that an obstruction
  is affecting the connectivity in a number of
  edges as shown in the figure.
• Give an estimate for the ditances dAB, dBC
  and dAC
• Answer
• dAB hAB x dhop 3 x 3 9
• dBC hBC x dhop 2 x 3 6
• dAC hAC x dhop 5 x 3 15

29
Questions

• Q4 Assuming accurate distance measurements
  between nodes, apply the trilateration technique
  to determine the SN coordinates (unknown) using
  the three BNs coordinates and the r distances
  (known). Let BN3 be the origin of the coordinate
  system.
• Answer

BN3 (0,0)
r3
SN (xs,ys)
r2
r1
30
Questions
31
References

• I. Stojmenovic, Handbook of Sensor Networks,
  Chapters 9 and 14, John Wiley Sons, 2005
• T. HE, C. HUANG, B. 11. Blum, J. A. Stankovic,
  and T. Abdelzaher, "Range-free localization
  schemes for large scale sensor networks, Proc.
  11obiCom'03, Sep. 2003, pp. 81-95.
• N. Bulusu, 1. Heidemann, and D. Estrin, "GPS-less
  low cost outdoor localization for very small
  devices," IEEE Personal Communications Magazine,
  vol. 7, 11ay. 2000, pp. 28-34.
• Boukerche, A. Oliveira, H.A.B. Nakamura, E.F.
  Loureiro, A.A.F. , "Localization systems for
  wireless sensor networks," Wireless
  Communications, IEEE , vol.14, no.6, pp.6-12,
  December 2007
• Sayed, A.H. Tarighat, A. Khajehnouri, N. ,
  "Network-based wireless location challenges
  faced in developing techniques for accurate
  wireless location information,"Signal Processing
  Magazine, IEEE , vol.22, no.4, pp. 24- 40, July
  2005

32
Thank you!

• Questions!