1. Genetic algorithms applied in routing protocols for wireless sensor networks
Ioana Apetroaei
Ionuţ-Alexandru Oprea
Bogdan-Eugen Proca
Laura Gheorghe
POLITEHNICA University of Bucharest
10th RoEduNet International Conference
June 23-25, 2011, Iasi, Romania

2. Contents Introduction & Context Protocol design
Genetic Algorithm overview
Testing environment
Results
Conclusions & Future Work
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

3. Introduction WSN Characteristics: Applications: Limited battery life
Low processing power
Random distribution
Medium-large coverage area
Applications:
Military surveillance
Healthcare monitoring
Industrial control
Environmental monitoring
Smart home
Military applications
Monitoring friendly forces, equipment and ammunition
Reconnaissance of opposing forces and terrain
Battlefield surveillance
Battle damage assessment
Nuclear, biological and chemical attack detection
Environmental applications
Forest fire detection
Biocomplexity mapping of the environment
Flood detection
Precision agriculture
Health applications
Tele-monitoring of human physiological data
Tracking and monitoring patients and doctors inside a hospital
Drug administration in hospitals
Home and other commercial applications
Home automation and Smart environment
Interactive museums
Managing inventory control
Vehicle tracking and detection
Detecting and monitoring car thefts
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

4. Network parameters 10 – 1000 nodes 20 – 30m (indoor)
70 – 100m (outdoor)
1 month – 3 years
128 bytes per packet
Base Station (BS)
Receives data from the network
High processing power
Unlimited power supply
Connected to other networks
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

5. Motivation Problem Solution
Uneven consumption of the energy of each node within the network
The network with a fixed topology becomes unusable when nodes closer to BS fail
Solution
Protocol for changing the topology of the network based on node energy levels in order to maximize the network lifetime
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

6. Protocol design (1) Initial topology
Each node determines the distances to its neighbours (RSSI)
Each node sends the distances to the BS
The genetic algorithm is based on these distances
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

7. Protocol design (2) Topology recalculation
Done by BS, after receiving energy levels from all nodes
Periodic
Topology = a spanning tree used for routing
Higher level in topology = more energy consumed by the node
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

8. Protocol design (3) - topology recalculation -
Scheduled – a timer on each node
Only packets with energy levels are sent; Easy to implement
Network flooding; Unrequested/delayed recalculations
Triggered – a node requests it
The topology is recalculated only when needed
BS broadcasts energy level request packets
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

9. Protocol design (4) - topology dynamics -1 2 3 1 4 5 6 3 2 1 7 8 9 6 5 4 4 5 6 9 7 8 7 2 9 3 8
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

10. Genetic Algorithm (1) Initial population Parent selection Crossing
Repair
Fitness
Mutation
Annealing
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

11. Genetic Algorithm (2) Chromosome generation All nodes are unmarked
BS is marked as visited
A random edge is added, having exactly one node marked (N-1 times)
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

12. Genetic Algorithm (3) Repair
Chromosome = parents array for the routing process
Two conditions must be met:
each node to be able to communicate with his parent, in the real topology
the resulting graph has to be a tree
If those two are not fulfilled, the algorithm returns to the crossing step
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

13. Genetic Algorithm (4) Fitness Fitness function for one chromosome
The new chromosome is accepted if it is better than one of its parents
If not, it still can be accepted with a probability that decreases with the generation number
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

14. Testing environment (1)
Technologies
C, nesC
TinyOS – operating system for WSNs
TOSSIM – simulator for TinyOS applications
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

15. Testing environment (2) - workload -
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

16. Results (1)
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

17. Results (2)
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

18. Results (3)
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

19. Future Work Transmit partial parents array associated with the subtree
One packet per node with its parent
The packet has to be fragmented and then reassembled at the destination in order to support larger networks
Improve the energy cosumption model
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

20. Conclusions Innovation related to the protocol design
Initial population
Fitness function
Topology recalculation
Simple routing protocol – the first node depletes after 14 rounds
Genetic routing protocol – after 18 rounds
The nodes work together – reduced standard deviation
RoEduNet 2011 – Genetic algorithms applied in routing protocols for wireless sensor networks

21. Genetic algorithms applied in routing protocols for wireless sensor networks
Ioana Apetroaei
Ionuţ-Alexandru Oprea
Bogdan-Eugen Proca
Laura Gheorghe
POLITEHNICA University of Bucharest
10th RoEduNet International Conference
June 23-25, 2011, Iasi, Romania