On SendingCoverage in Sensor Networks

1
On Sending-Coverage in Sensor Networks

• Guoliang Xing
• Department of Computer Science and Engineering
• Washington University in St.Louis

2
Motivation

• Challenge
• achieve long life time on limited energy
• Assumptions
• Large scale, ad hoc node deployment
• Dense networks fault tolerance, long system life
• Approach
• Use active nodes to provide sufficient service
• Schedule unnecessary nodes to sleep

3
Sufficient Service

• Sensing coverage
• Sensor modality acoustic, magnetic, seismic ..
• Applications localization, object tracking,
  detection, structural monitoring
• Communication
• K-Connectivity the network is still connected if
  (K-1) nodes fail
• Routing quality hop count

4
Limitations of Existing Protocols

• Treat communication and coverage in isolation
• Connectivity only ASCENT, SPAN, AFECA, GAF,
• Coverage only exposure, Ottawas protocol,
• Density PEAS
• Lack flexibility fixed degree of coverage
  every point in a region is covered by K nodes

5
Outline

• Coverage Configuration Protocol (CCP) (Sensys03)
• Different degree of coverage
• Allows integrated coverage connectivity
  configuration
• Impact of coverage on communication quality
• Network connectivity
• Routing performance (Mobihoc04)
• Probabilistic coverage (IPSN 04)
• Co-Grid detection-based coverage protocol

6
Disc Comm./Coverage Model

• A point p is covered by node v if pv lt Rs
• Rs Sensing range
• K-coverage every point in a region is covered by
  at least K nodes
• Nodes u and v are connected if uv lt Rc
• Rc Communication range

7
A Sufficient Condition for K-Coverage

• Theorem A convex region B is K-covered if all
  the intersection points inside B are K-covered
• Intersection points among sensing circles
• Intersection points between circles and Bs
  boundaries
• A node is eligible to become active iff there
  exists at least one intersection point inside its
  sensing circle that is not K-covered

on?
Active nodes
Sleeping nodes
Intersection point
8
Coverage Configuration Protocol (CCP)

• Sleeping node periodically wake up
• Listen to neighbors location beacons and
  announcements
• If eligible
• Set a random timer
• Turn active and broadcast JOIN if it is still
  eligible when timer expires
• Active node
• Listen to neighbors location beacons and
  announcements
• If ineligible
• Set a random timer
• Broadcast WITHDRAW and sleep if still ineligible
  when timer expires

9
Integrated Coverage and Connectivity Configuration

• If Rc/Rs ? 2 ?
• Theorem A K-covered network is also K-connected
• If Rc ? 2Rs, only need to configure coverage
• Solution Coverage Configuration Protocol (CCP)
• If Rc lt 2Rs, must address both coverage and
  connectivity.
• Solution CCP SPAN

10
SimulationCoverageConnectivity (Rc 1.5Rs)
SPAN
CCP
SPANCCP

• Combination of SPAN CCP is necessary for
  desired coverage and connectivity when Rc lt 2Rs

11
Outline

• Coverage Configuration Protocol (CCP) (Sensys03)
• Different degree of coverage
• Allows integrated coverage connectivity
  configuration
• Impact of coverage on communication quality
• Network connectivity
• Routing performance (Mobihoc04)
• Probabilistic coverage (IPSN 04)
• Co-Grid detection-based coverage protocol

12
Metric Network Dilation

• Network dilation

• Shortest hop count between u,v in G(V,E)

• When is the hop count of the
  shortest path chosen by a routing algorithm R, Dn
  characterizes the performance of R

13
Preliminaries Voronoi Diagram and Delaunay
Triangulation (DT)

• Voronoi diagram of a set of nodes V
• The partition of the plane into V Voronoi cells
• A point p lies in the Voronoi cell of node s iff
  s is closer to to p than any other node in V
• Delaunay Triangulation (DT) dual graph of
  Voronoi diagram

14
Analysis based on DT

• Known result DT has low dilation property
• Theorem when Rc/Rs2, DT is a sub-graph of a
  sensing-covered network
• ?Network dilation is 4.84, i.e., there exists a
  path less than hops between u and v
• Too high when Rc/Rs gtgt 2

15
Greedy Forwarding (GF) Routing

• Always choose the neighbor closest to destination
  as the next hop
• Simple and highly efficient
• Fail in presence of network voids
• Need complex routing mode to recover, e.g., face
  routing travel around network voids

16
GF Always Succeeds in Sensing-covered Networks

• Theorem Max hop count between uv is
• Too conservative when Rc/Rs approaches 2

Advance at least Rc-2Rs
Destination node
Routing node
Always find a next hop due to sensing coverage
17
Motivation for New Approach

• Greedy forwarding (GF)
• Network Dilation is Rc/(Rc-2Rs)
• Too conservative when Rc/Rs approaches 2
• DT Dilation
• Network Dilation is 4.84
• Too conservative when Rc/Rs gtgt 2
• Combining them together is better

18
Bounded Voronoi Greedy Forwarding (BVGF)

• Leverage the greedy nature of GF and constant
  dilation property of DT
• Routing decisions are made based on local Voronoi
  diagram
• Achieve low dilation for all Rc/Rs

19
BVGF

• Next-hop candidates the neighbors whose Voronoi
  cells is intersected by uv
• Next hop the candidate closest to destination

source node
destination node
routing node
next hop
20
Asymptotic Network Dilation under BVGF

• Theorem

• Worst case dilation 4.62
• i.e., BVGF always finds a routing path less than
  hops between u and v

21
Simulations Network Dilation
BVGF achieves low dilation under all Rc/Rs
Both GF and BVGF approach optimal when Rc/Rs is
large
Measured dilations under GF/BVGF is low
22
Outline

• Coverage Configuration Protocol (CCP) (Sensys03)
• Different degree of coverage
• Allows integrated coverage connectivity
  configuration
• Impact of coverage on communication quality
• Network connectivity
• Routing performance (Mobihoc04)
• Probabilistic coverage (IPSN 04)
• Co-Grid detection-based coverage protocol

23
Motivation

• Deterministic Sensing Model
• Hard artificial boundary between covered and
  not covered
• Probabilistic coverage
• Coverage probability decays with the distance

Network Topology with Sensing Coverage
24
Problem Formulation

• Minimize the number of active sensors under the
  coverage constraint
• A geographic region A is covered if
• PD(x,y) Fused detection probability of all
  active sensors at point (x,y)
• PF False alarm rate

25
Data Fusion

• Decide local fusion groups
• Design a set of decision rules in a fusion group
• Majority rule is used

S1 sensors with decision 1 S0 sensors with
decision 0
O(2n) combinations
26
Centralized Coverage Algorithm

• while (PDmin lt ß)
• Find point p(x,y) with min PD(x,y)
• Activate the sensor closest to p
• ?????
• Fusing all sensor decisions at single fusion
  center
• Ignore signal locality
• High computational cost

O(2n) n number of active sensors
27
Se-Grid Coverage Algorithm based on Separate
Grids

• Divide the region into multiple grids
• Fusion center in each grid runs the centralized
  algorithm

• Configuration time is reduced via parallel
  processing
• Data fusion is restricted within each grid ?
  redundant active sensors

28
Go-Grid Coverage Algorithm with Inter-grid
Coordination

• Overlapping grid layout
• Inter-grid coordination is needed

2 X 2 Grids
G(1,1)
G(2,1)
G(2,2)
G(1,2)

• Each grid has 4 squares and each square belongs
  to (up to) 4 grids

29
Performance Number of Active Sensors
Co-Grid is competitive with Central when grid
width gt ¼ region width
30
Conclusions

• Integrated coverage connectivity configuration
• Analysis of network connectivity and routing
  performance
• Better geographic routing algorithm
• Detection-based probabilistic coverage model
  protocol

31
Collaborators References

• Collaborators Dr. Chenyang Lu (advisor)
  Dr.Robert Pless Dr. Qingfeng Huang Xiaorui
  Wang Yuanfang Zhang Dr. Joseph OSullivan Dr.
  Chris Gill
• References
• On Greedy Geographic Routing Algorithms in
  Sensing-Covered Networks, Guoliang Xing Chenyang
  Lu Robert Pless Qingfeng Huang,MobiHoc 2004
• Co-Grid An Efficient Coverage Maintenance
  Protocol for Distributed Sensor Networks,
  Guoliang Xing Chenyang Lu Robert Pless Joseph
  A. O'Sullivan IPSN'04
• Integrated Coverage and Connectivity
  Configuration in Wireless Sensor Networks,
  Xiaorui Wang Guoliang Xing Yuanfang Zhang
  Chenyang Lu Robert Pless Christopher D.
  Gill,SenSys'03