Locate More Sensors with Fewer Anchors in Wireless Sensor Networks

1
Locate More Sensors with Fewer Anchors in
Wireless Sensor Networks

• Hui Ling and Taieb Znati, Department of Computer
  Science, University of Pittsburgh
• hling, znati_at_cs.pitt.edu

2
Outline

• Location discovery in Wireless Sensor Networks
  (WSNs) and Challenges
• Locate more sensors using Out-of-Range
  information
• Simulation results
• Conclusion

3
Why Location Is Needed in WSNs?

• Collected data is typically interpreted in
  reference to location information
• To track moving objects inside the network
• To locate events occurred in the network

4
Location Discovery in WSNs

• Range based schemes
• Distance estimation
• Received Signal Strength Indicator (RSSI)
• Time of Arrival (TOA)
• Time Difference of Arrival (TDOA)
• Angle of Arrival (AOA)
• Distance combining
• Multi-lateration (Triangulation)
• Range free schemes Centroid, DV-HOP, APIT

5
Multi-lateration Example
What if unknown nodes can not gain sufficient
information of reference nodes during location
discovery?
Reference node
Unknown node
Unresolved node
6
Our Approach

• Integrate Out-of-Range information into range
  based schemes to
• Eliminate location ambiguities of sensor nodes in
  sparse networks
• Reduce the number of anchor nodes in WSNs for
  location discovery

7
Outline

• Location discovery in Wireless Sensor Networks
  (WSNs) and Challenges
• Locate more sensors using Out-of-Range
  information
• Simulation results
• Conclusion

8
Out-of-Range Information
N1
N2
r
N1 can not hear from N2 ? Distance(N1,N2) gt r
9
Use Out-of-Range for Location Estimation Case 1
P
If the unknown node is out of R3s range, then
the unknown node can only be at P since the other
potential location P is within R3s range.
R3
R2
R1
P
Reference node
Unknown node
10
Use Out-of-Range for Location Estimation Case 2
Unknown node U1 can not be here if U1 is out of
U2s range.
U2
As a result, U1 must be here.
U1
Reference node
Unknown node
11
Use Out-of-Range for Location Estimation Case 3
Unknown node U1 can not be here if U1 is out of
U2s range.
Therefore, U1 must be here.
U2
U1
Reference node
Unknown node
12
Location Estimation using Out-of-Range Information

• Maintain Out-of-Range information
• Neighbors are discovered through a periodical
  message exchange protocol
• Non-neighboring nodes are out of each others
  transmission range
• Location discovery
• Unknown nodes collect distance and location
  information of neighboring reference or resolved
  nodes
• If fewer than 3 neighboring reference nodes
  exist, unknown nodes send out help messages to
  all h-hop neighbors
• Upon receiving a help message from U, node N
  determines if its location (actual or estimation)
  information is useful. If so sends the
  information back to U
• After gaining sufficient useful out of range
  information from k-hop neighbors, U can determine
  its location

13
Location Discovery using Out-of-Range
Case1
R1
U1
Reference node
Unknown node
14
Location Discovery using Out-of-Range
Case2
Case1
Help
Help
Reference node
Unknown node
15
Outline

• Location discovery in Wireless Sensor Networks
  (WSNs) and Challenges
• Locate more sensors using Out-of-Range
  information
• Simulation results
• Conclusion

16
Simulation Results

• Network setup
• In an area of 1000m1000m
• Transmission range is 250m
• The number of nodes vary from 16 to 48
• Performance evaluation
• Effect of h, the number of hops to send
  LOCATION_HELP message
• Effect of network density on the number of
  resolved nodes after location discovery
• Location estimation error

17
Effect of h
18
Effect of Average Node Degree
19
Number of Anchors to Resolve All Sensors in the
Network
20
Location Estimation Error
21
Location Estimation Error
22
Conclusion

• Out-of-Range information can be integrated into
  multi-lateration based schemes to
• Reduce the number of anchors to discover
  locations of sensor nodes in dense WSNs
• Locate more sensors in sparse WSNs
• Thank you! and Questions?

23
Network Density