Cut Detection in wireless sensor networks.

1. By:
Pulkit Chhabra
10103644

2. Topics for discussion
 Introduction
 Sensor network
 Architecture of sensor network.
 Applications of sensor network.
 Challenges of sensor network.
 Cuts in sensor networks.
 Detection of cuts.
 Distributed cut detection algorithm.
 Conclusion.
 Future aspects.
 References.

3. Sensor Network:
It is a group of specialized transducers with a
communication infrastructure intended to monitor
and record conditions at diverse location, it may be
temperature, pressure humidity wind direction and
speed etc…
It consists of multiple detection station called
nodes, which is small, lightweight and portable.

4. Contd…

5. Contd…

6. Architecture of WSNs
 Transducer generates electrical signals.
 Microcomputer processes and stores
Sensor output.
 Transceiver receives the
commands from central
computer, transmits data
to that computer.
 Power from battery.

7. Applications
 Industrial monitoring.
 Machine health monitoring.
 Data logging.
 Industrial sense and control applications.
 Automated and smart homes.
monitoring the activities performed in a smart
home is achieved using wireless sensors embedded
within everyday objects forming a sensor network.
 Medical device monitoring.

8. Contd…
 Monitoring of weather conditions.
 Air traffic control.
 Robot control.
 Air pollution monitoring.
 Forest fire detection.
 Natural disaster prevention.

9. Challenges
 One of the challenge in the successful use of WSNs
come from limited energy of the individual sensor
nodes.
 WSNs consist of large number of small, low cost
sensor nodes distributed over large area.
 WSNs have emerged as a promising new technology
to monitor large regions.
 Node failure is expected to be quite common. this is
true for sensor networks deployed in harsh and
dangerous fields such as forest fire monitoring and in
defense applications.

10. Cuts in sensor networks
 WSNs can get separated into multiple connected
components due to failure of some of its nodes, which
is called “cut”.
 So here we consider the problem of detecting cuts.
 Two nodes are said to be disconnected if there is no
path between them.

11. Contd…

12. Contd…
 In the figure, filled circles represents active nodes and
unfilled circles represent failed node.
 Solid lines represents the edges, and dashed lines
represent the edges that existed before the failure of
node.
 the hole in (d) is indistinguishable from the cut in (b)
to nodes that lie outside the region R

13. Detection of cuts
 Nodes that detect the occurrence and approximate the
locations of the cuts can then alert the source node or base
station.
 If the node were able to detect the occurrence of cut it could
simply wait for the network to be repaired and eventually
reconnected, which saves on-board energy of multiple nodes
and prolongs their lives.
 The ability to detect cuts by both the disconnected node and
source node will lead to increase in the operational lifetime
of the network as a whole.

14. Distributed cut detection algorithm
 This algorithm allows each node to detect
DOS(Disconnection frOm Source) events and subset of node
to detect CCOS(connected but cut occurred somewhere)
events.
 The algorithm is distributed and asynchronous.
 It involves only local communication between neighboring
nodes and is robust to temporary communication failure
between node pairs.
 A key component of the DCD algorithm is a distributed
iterative computational step through which the nodes
compute their electrical potentials.

15. Distributed cut Detection
Definition and problem statement
 Time is measured with a discrete counter k=∞,…., -
1,0,1,2,…..
 We model the sensor network as a time varying graph
𝒢(𝑘) = (𝑉(k )ℰ(k)), whose node set V(k) represents the
sensor nodes active at time k and the edge set ℰ(𝑘)
consist of pairs of nodes (u,v) such that nodes u and v can
directly exchange the message between each other at
time k.
 By an active nodes we mean a nodes that has not failed
permanently .

16. Contid…
 All graphs considered here are undirected i.e (i,j)=(j,i).
 The neighbors of node i is the set 𝑁𝑖 of nodes connected to i
i.e. 𝑁𝑖 = {𝑗|(𝑖, 𝑗) ∈ ℰ}.
 The numbers of neighbors of i, |𝑁𝑖 𝑘 |, is called its degree,
which is denoted by di.
 A path from i to j is sequence of edges connecting i and j.
 A graph is called connected if there is a path between every
pair of nodes.
 A component Gc of the graph is maximal connected sub
graph of G.

17. Contid…
 A cut is formally defined as the increase of the
number of components of a graph due to the failure of
subset of nodes.
 The number of the cuts associated with a cut event is
the increase in the number of components after the
event.

18. Contid…
 The problem we seek to address is two fold
 First we want to enable every node to detect if it is
disconnected from source.
 Second we want to enable nodes that lie close to the cuts
but are still connected to the source.

19. Contid…
 The DCD algorithm is based on the following electrical
analogy.
 Imagine the wireless sensor network as an electrical circuit
where the current is injected at the source node and
extracted out of a common fictitious node that is connected
to every node of sensor network.
 Each edge is replaced by the 1 Ω resistor.
 When a cut separates certain nodes from source node, the
potential of each of those nodes becomes 0, since there is no
current injection into their components.
 The potential is computed by an iterative scheme which only
requires periodic communication among the neighboring
nodes.

20. Contid…
DOS detection
• When node u is disconnected from the source, we say that a
disconnected from source event has occurred for u.
• The algorithm allows each node to detect DOS events.
• The nodes use the computed potentials to detect if DOS
event have occurred .
• The approach here is if the state is close to 0 then the node is
disconnected from the source , otherwise not.

21. Contid…
 DOS detection part consist of steady state detection,
normalized state computation and
connection/separation detection.

22. Contid…
CCOS detection
 When a cut occurs in the network that does not separate a
node u from the source node, connected but a cut occurred
somewhere (CCOS) event has occurred for u.
 Detection of CCOS events by the nodes close to a cut, and
approximate location of cut means location of one or more
active nodes that lie at the boundary of cut and that are
connected to source.

23. Contid…
 To detect the CCOS event the algorithm uses the fact
that the potential of nodes that are connected to
source node changes after the cut and also using probe
messages.
 Probe messages that are initiated by the certain nodes
that encounter failed neighbors, and are forwarded
from one node to another node in a way that if short
path exist around a hole created by node failures, the
message will reach initiating nodes.

24. Contid…
 Each probe message contains the following
information:
 A unique source ID
 Source node ID S
 Destination node,
 path traversed, and
 angle traversed by probe message.

25. Conclusion
 The DCD algorithm we propose here enables every
node of a wireless sensor network to detect
disconnected from source event if they occur.
 Second it enables the subset of nodes that experiences
CCOS event to detect them and estimate the
approximate location of the cut in the form of a list of
active nodes that lie at the boundary of the cut/hole.

26. Future Aspect
 Application of DCD algorithm to detect the node
separation and reconnection to the source in mobile
networks.

27. references
 http:// www.liveieeeprojects.com
 http://www.ijsrp.org/research-paper-1212/ijsrp-p1203.pd
 http://www.faculty.cs.tamu.edu/stoleru/papers/prabir12c
ut.pdf - United States
 www.cs.ucsb.edu/~suri/psdir/sentinel.pdf
 www.slideshare.net/.../cut-detection-in-wireless-sensor-
networks
 http://en.wikipedia.org/wiki/Wireless_sensor_network

28. THANK YOU