Energy Aware Routing Protocol for Sensor Networks

1
Energy Aware Routing Protocol for Sensor Networks

• Sheetal Agarwal

2
Agenda

• Introduction to Sensor Networks
• Sensor Network Vs Fixed and Ad-Hoc Networks
• Routing Protocols for Sensor Networks
• Proposed Routing Protocol
• Performance Metrics

3
Introduction to Sensor Networks

• What is a Sensor?
• A sensor is a device the produces a measurable
  response to a change in physical condition such
  as temperature or chemical condition such as
  concentration
• Types of Sensors - Temperature, Pressure,
  Humidity, Image etc

4
Sensor Network Architecture
Sink
Internet/Satellite
Sensor Nodes
Task Manager
User
5
Sensors
EmbedSense - Wireless Sensor and Data Acquisition
System
Berkeley Mote
6
Applications of Sensor Networks

• Monitoring Applications
• Military
• Security
• Intrusion Detection
• Habitat Monitoring
• Environment Observation and Forecasting

7
Sensor Vs Ad-Hoc Networks

• Network topology is not fixed
• Power is an expensive resource in these networks
• Nodes are connected by wireless links

8
Sensor Vs Ad-Hoc Networks

• Large number of sensors
• Addressing scheme for sensor nodes
• Sensor network used for data gathering
• Ad - Hoc network used for distributed
  computing

9
Sensor Vs Ad - Hoc Networks

• Data flows from multiple sources to a single
  destination
• Redundancy in data traffic
• Sensor nodes are prone to failure

10
Classification of Routing Protocols

• Data Centric Protocols
• SPIN , Directed Diffusion
• Hierarchical Protocols
• LEACH , TEEN
• Location Based Protocols
• GAF , GEAR

11
Data Centric Routing

• Address Centric Routing
• Finding short routes between pairs of addressable
  end nodes
• Data Centric Routing
• Perform in-network consolidation of redundant
  data while routing from source to the sink

12
Data Aggregation

• Methods of Aggregation
• Duplicate suppression
• Aggregate functions like Avg,Min,Max etc
• Data Aggregation Trees
• Center At Nearest Source
• Shortest Path Tree
• Greedy Incremental Tree

13
Data Distribution

• Flooding
• Data packet is broadcasted by each node to all
  its neighbors
• Causes Implosion, Overlap Resource Blindness
• Gossiping
• Each receiving node sends the packet to a
  randomly selected neighbor
• No Implosion but takes more time

14
Hierarchical Protocols

• When sensor density increases single tier
  networks cause
• Gateway overloading
• Increased latency
• Large energy consumption
• Clustered Network allow coverage of large area of
  interest and additional load without degrading
  the performance

15
Hierarchical Protocols

• Hierarchical routing
• Uses Multi - hop communication within a cluster
• Performs data aggregation and fusion on data to
  reduce number of transmitted messages to the sink
• Maintain the energy reserves of nodes efficiently
• Example - LEACH, PEGASIS

16
Location Based Protocols

• Location information can be used to
• Find shortest path to the sink
• Form a virtual grid and keep only few nodes
  active at a time
• Example
• GAF
• GEAR
• SPAN

17
Determining Location

• Location of a node can be determined using
• Global Positioning System
• Ultrasonic Systems using trilateration
• Beacons
• Location based protocols assume that each node
  knows its location in the network

18
GAF(Geographic Adaptive Fidelity)

• Forms a virtual grid of the covered area
• Each node associates itself with a point in the
  grid based on its location
• Nodes associated with same point in grid are
  considered equivalent
• Some nodes in an area are kept sleeping to
  conserve energy
• Nodes change state from sleeping to active for
  load balancing

19
GAF

• A node remains active for time Ta
• Ta of a node in the grid is broadcasted to other
  equivalent nodes
• The sleeping time of a node is adjusted depending
  on Ta
• In the discovery state each node broadcasts
  discovery messages periodically (Td)

20
State Transition for GAF
Sleeping
After Ts
After Td
Discovery
Active
After Ts
21
Routing in GAF
22
GAF

• Not very scalable. As the network size increases
  distance to the base station increases
• Only the active nodes sense and report data.
  Hence data accuracy is not very high.

23
GEAR (Geographically and Energy Aware Routing)

• Queries
• Contain location information
• Disseminated to only the specified region of the
  network
• Neighbors are selected probabilistically to
  forward the query to the target location
• Query is flooded only in the target region

24
GEAR

• Each node maintains a neighbor table
• Energy levels and locations of each neighbor
• Cost to transmit to each neighbor
• Packet is forwarded to neighbor with smallest
  cost

25
GEAR
26
GEAR

• Not Scalable
• All nodes are active even though only a part of
  the network is queried

27
Enhanced - GEAR (EGEAR)

• Many applications require data on demand
• Not all nodes need to be active and sensing data
  continuously
• Few nodes can be active to keep the network
  connected
• Remaining nodes are activated when data is
  required

28
Creating a Virtual Grid

• Use location information to create a virtual grid
• All nodes in a grid are equivalent
• Only one node from a grid point is active at a
  time
• Each node in a grid point is within the radio
  range of nodes in adjacent grids
• Once the virtual grid is created, hierarchical
  clusters of these grids are created

29
Cluster Formation

• Let N be the number of clusters that are formed.
  For a given round a node becomes cluster head if
• Each node chooses a random number between 0 and
  1. Node becomes cluster head if number less than
• T(n) P/((1-P (r mod 1/P))) if n E G
• 0 otherwise
• P - desired of cluster heads, r round, G set of
  nodes
• that have not been cluster head for last 1/P
  rounds

30
Cluster Formation

• Each cluster head broadcasts its location
  information
• Nodes join associate themselves with the closest
  cluster head and inform the cluster head of their
  decision

31
Query Dissemination

• Each query consists of location information
• Base station forwards the query to the cluster
  head closest to the target region
• Each cluster head forwards the query to the
  closest cluster head
• Cluster head activates all nodes in the in the
  target region

32
Data Collection

• Each node transmits data to the cluster head
• The data is aggregated by the cluster head
• The aggregated data is then sent to the base
  station using reverse path

33
(No Transcript)
34
EGEAR

• Keeps minimum number of nodes active in the
  network
• Reduces the number of query packets disseminated
• Improves quality of data delivered
• Scalable
• Suitable for demand driven applications

35
Testing

• Use Network Simulation tools like Glomosim
• Test with networks of 400-5000 nodes
• Consider static sensor networks only
• Determine optimal number of clusters for a given
  network size

36
Performance Metrics

• Average Energy Consumption
• Impact of localization errors
• Energy and time expended for cluster formation
• Network lifetime

37
Questions?