1. 2008/2/191 Customizing a Geographical Routing Protocol for Wireless Sensor Networks Proceedings of the 2005 11th International Conference on Information Technology: Coding and Computing Jian Chen Department of Computer Science Texas A&M University College Station, TX 77843 Email: jchen@cs.tamu.edu Yong Guan Department of Electrical & Computer Engineering, Iowa State University Ames, IA 50011 Email: guan@ee.iastate.edu Udo Pooch Department of Computer Science Texas A&M University College Station, TX 77843 Email: pooch@cs.tamu.edu

2. 2008/2/192 Outline 1. Introduction 2. Algorithm and Implementation 3. Simulation results and Evaluation 4. Discussion 5. Future work

3. 2008/2/193 Introduction(1/5)  Greedy Perimeter Stateless Routing (GPSR)  GPSR is designed under the assumption of symmetric wireless links  In sensor networks, packet destinations are often marked with locations instead of identifiers like IP addresses  A simple beaconing algorithm  Greedy mode and Perimeter mode

4. 2008/2/194 Introduction(2/5)  Greedy Mode

5. 2008/2/195 Introduction(3/5)  Perimeter Mode  Right-hand Rule

6. 2008/2/196 Introduction(4/5)  Perimeter Mode

7. 2008/2/197 Introduction(5/5)  Greedy mode or Perimeter mode Greedy modePerimeter mode greedy fails have left local maximagreedy worksgreedy fails

8. 2008/2/198 Algorithm and Implementation(1/6)  On demand GPSR (OD-GPSR)  Usage of symmetric and asymmetric wireless links  Soliciting Beacons from Neighbors broadcast a one-hop beacon-request beacon is a one hop broadcast packet or a one hop unicast packet power save and asymmetric link detect  Greedy Forwarding and Right-Hand Rule  Boundary Problem

9. 2008/2/199 Algorithm and Implementation(2/6)  On demand GPSR (OD-GPSR)  symmetric wireless links

10. 2008/2/1910 Algorithm and Implementation(3/6)  On demand GPSR (OD-GPSR)  asymmetric wireless links  When a neighbor receives notification of delivery failure for unicast beacon packets (we assume the MAC layer has such capability) for several times, the neighbor believes the link is unidirectional and then sends a special unidirectional notification beacon via a local broadcast packet which defines the maximum hops allowed

11. 2008/2/1911 Algorithm and Implementation(4/6)  On demand GPSR (OD-GPSR)  asymmetric wireless links

12. 2008/2/1912 Algorithm and Implementation(5/6)  On demand GPSR (OD-GPSR)  Boundary Problem  Three step 1.detection of a packet with outside target location 2.collect boundary information 3.inform all border nodes of the collected boundary information

13. 2008/2/1913 Algorithm and Implementation(6/6)  On demand GPSR (OD-GPSR)  Boundary Problem

14. 2008/2/1914 Simulation results and Evaluation(1/6)  GPSR vs OD-GPSR  Average energy consumption  Packet delivery success Rate  Average delay  Since the latter works under the assumption of known boundary, we limit traffic destinations inside the network topology for the convenience of comparison.

15. 2008/2/1915 Simulation results and Evaluation(2/6)  We simulated OD-GPSR in ns-2  802.11 physical and MAC layer  256 nodes are randomly deployed in a 256m by 256m rectangle area  The radio range is changed to 40 meters to make it closer to the real situation.  GPSR-bint5 means GPSR with beacon interval of 5 seconds and ODGPSR-bint5 means OD- GPSR with beacon interval of 5 seconds

16. 2008/2/1916 Simulation results and Evaluation(3/6) Average energy consumption

17. 2008/2/1917 Simulation results and Evaluation(5/6) Average delay

18. 2008/2/1918 Simulation results and Evaluation(4/6) Packet delivery success Rate

19. 2008/2/1919 Simulation results and Evaluation(6/6) Average energy consumption - all unicast beacon

20. 2008/2/1920 Discussion(1/1)  As GPSR, OD-GPSR guarantees the delivery of packets if it is applied to an environment where all nodes have the same transmission range, but performs better than GPSR in terms of energy efficiency and data delivery rate at the cost of a little bit more delay.  It is applied to an environment where link asymmetry exists.

21. 2008/2/1921 Future work(1/1)  how to guarantee the delivery of packets under situations where non-uniform transmission ranges exist  we will improve our protocol to decrease the delay