Non-uniform Grid-based Coordinated Routing

1. Non-uniform Grid-based Coordinated Routing PriyankaKadiyala Major Advisor: Dr. Robert Akl Department of Computer Science and Engineering

2. Outline • Research objective • Overview of sensor networks • Related work • Motivation • Non-uniform grid-based coordinated routing protocol • Simulations and results

3. Overview of Sensor Networks • Ad hoc networks of tiny battery powered sensor nodes capable of sensing, processing and communicating data. • Applications - Video surveillance, traffic monitoring, environmental monitoring, structure and system health monitoring in buildings and aircraft interiors. • The main source of energy is battery, no external supply of power - major constraint is energy available.

4. Research Objective To increase the lifetime of the sensor network by using non-uniform grid based routing for the case of random node deployment.

5. Overview of Sensor Networks

6. Overview of Sensor Networks Protocols for WSNs • Flooding, Gossiping, SPIN, LEACH, PEGASIS, Directed Diffusion, and GEAR • Energy efficient protocols that allow nodes to be put to sleep are GAF, SPAN, STEM, ASCENT, CEC, AFECA and GBCR.

7. Related Work Flooding : • In flooding every node that receives a packet broadcasts it to its neighbors. If the node receives the packet for the first time, it is stored in the buffer. If it is a redundant packet, it is discarded.

8. Flooding Simulation

9. Related Work Geographic Adaptive Fidelity: • A virtual grid is proposed with only one node active at a time in each grid. • Other nodes save energy by turning their radios off, or by entering sleep mode. • Each node in GAF has three states: sleeping, discovery and active states respectively.

10. Related Work Span: • Forms a backbone network of active nodes that participate in routing. • A node in Span can only be in two states: coordinator and a non-coordinator. • A node volunteers to be the coordinator if two of its neighbors fail to communicate with each other, either directly or through another coordinator.

11. Related Work Grid-based Coordinated Routing: • Combines flooding, GAF and Span. • Network is partitioned into square shaped grids. • In each grid, one node participates in routing while other nodes are put to sleep to conserve energy.

15. Motivation • To save energy by radio range adjustment, dividing the network into sections of different grid sizes based on a range-traffic relationship has been proposed. • Our work is motivated from the concept of non-uniform grid sizes across the network using coordinated routing.

16. Non-uniform Grid-based Coordinated Routing Protocol • The entire test area is divided into grids. • Estimate the grid size to ensure proper connectivity between two coordinator nodes in adjacent grids. • A coordinator node is elected in each grid to participate in routing. • Energy depletion of nodes is taken into account for load balancing in the network.

17. Estimating the Grid Size • To ensure connectivity and efficient usage of node energy, the grid size should neither be too large nor too small.

18. Estimating the Grid Size (Contd.) • The amount of energy that is required to establish a link between two nodes is proportional to the distance between the two nodes raised to a constant power, called the path loss exponent, n . • If S is the receiver sensitivity, the communication link between the two nodes leads to a successful transmission between the nodes if the power of the received signal is greater than S.

19. Estimating the Grid Size (Contd.) Rn r 2r We define an upper bound on the grid size as 200 m and consider a lower bound of 100 m.

20. Non-uniform grid structures Designing the grid structures: • Areas of high node density can be used efficiently for a grid size of 200 m. • Areas of low node density require a grid size of 100 m. • Random node placement implies sparsely and densely populated areas, therefore requiring a non-uniform distribution of grid size.

21. Types of non-uniform grids • Source non-uniform grid structure : suitable for low density around the source node and high density around sink node. • Sink non-uniform grid structure : suitable for high density around the source node and low density around the sink node. • Alternating non-uniform grid structure : suitable for random node placement across the network.

22. Source non-uniform grid structure

23. Sink non-uniform grid structure

24. Alternating non-uniform grid structure

25. Coordinator node election • Each node has a randomly assigned ID. • From each grid, the node with maximum node ID is the coordinator node . • To distribute load across the coordinator nodes in a fair manner, load balancing is employed.

26. Load Balancing • If coordinator node energy > 25% of battery life, node rank = node rank +1 • If the energy < 25% of battery life, node rank = node rank + 2 • For each grid, the current coordinators are replaced with lower ranked nodes.

29. Simulations and results Assumptions: • Energy consumption by nodes is assumed as Idle:transmit:receive = 1:2:1.5 • Test area is assumed to be replicating an actual sensor field of size 1000 m in the x-direction and 1000 m in the y-direction. • Position of nodes deployed is assumed to be the same for all grid structures.

32. Network Parameters

33. Results Metrics: • Normalized energy • Network lifetime Graphs: • Network lifetime graph. • Energy depletion graph.

34. Network Lifetime Graph

37. Energy Depletion Graph

41. Conclusion • Different non-uniform grid structures provide different levels of energy savings and network lifetime. • For random node deployment, using a non-uniform grid structure of alternating small and large grid size improves network lifetime over a uniform grid structure.

43. Future Work • Implementation on actual motes. • Mobility of nodes. • Irregular distribution of nodes.

44. Thank you Questions ?