Adaptive Triangular Deployment Algorithm for Unattended Mobile Sensor Networks

1. Adaptive Triangular Deployment Algorithm for Unattended Mobile Sensor Networks Ming Ma, Yuanyuan Yang IEEE Transactions On Computers 2007 Suho Yang (September 4, 2008)

2. Contents • Introduction • Assumption • Ideal Node Layout for Maximum Coverage • Two Triangular algorithms • Basic Triangular Algorithm • ATRI : Adaptive Triangular algorithm • Performance Evaluation • Conclusion

3. Introduction • Wireless Sensor Networks • Applications: military, environmental, health, home, commercial, … • Strong points • Small in size • Low cost → can be densely deployed • Weak points • Low computational capacities and memory • Short communication range • Low power consumption requirement: the most important metric • Other features • Wireless communication • Collaborative effort • Fault tolerance

4. Introduction • Deployment problem for WSN • Problem 1: “Where to place sensors?” • To maximize sensing coverage (using a certain number of sensors) • =To minimize coverage gaps and overlaps • → Propose the ideal node layout for maximum coverage • Problem 2: “How to move sensors?” • To minimize the total energy consumption to move sensors • → Propose a distributed greedy heuristic algorithm

5. Assumption • Assumptions • All sensors are mobile sensors • All sensors have the same capacities (=sensing range, energy, …) • Omni-directional sensing • No global information • No location-awareness • Each sensor only estimate the relative locations to neighbors • Initially, all sensors are randomly deployed

6. Problem 1: “Where to place sensors?”- Ideal Node Layout for Maximum Coverage

7. Ideal Node Layout for Maximum Coverage • The optimal node layout for the maximum no-gap coverage • Equilateral triangulation

8. Ideal Node Layout for Maximum Coverage • Proof • The maximum value of can be obtained when • In this case, the lengths of all three edges =

9. Problem 2: “How to move sensors?”- Basic Triangular Deployment Algorithm

10. Basic Triangular Deployment Algorithm • Main idea • Each node divides the transmission circle into six sectors • And adjusts the relative distance to its one-hop neighbors in each sector separately

11. Basic Triangular Deployment Algorithm • Adjusting the distance between neighbors • Notation where r = sensing range

12. Basic Triangular Deployment Algorithm • Adjusting the distance between neighbors (cont.)

13. Basic Triangular Deployment Algorithm • Adjusting the distance between neighbors (cont.)

14. Basic Triangular Deployment Algorithm • Adjusting the distance between neighbors (cont.)

15. Basic Triangular Deployment Algorithm • Algorithm

16. Basic Triangular Deployment Algorithm • Snapshots of the execution

17. Basic Triangular Deployment Algorithm • Reducing node oscillation • (1) Distance threshold strategy • Constant threshold • Variable threshold

18. Basic Triangular Deployment Algorithm • Reducing node oscillation • (2) Movement state diagram strategy

19. Problem 2: “How to move sensors?”- ATRI: Adaptive Triangular Deployment Algorithm

20. ATRI: Adaptive Triangular Deployment Algorithm • Additional consideration • 1. Avoiding obstacles and boundaries • 2. Non-uniform deployment

21. ATRI: Adaptive Triangular Deployment Algorithm • 1. Avoiding obstacles and boundaries • Detect them with an ultrasonic obstacle-detecting module • And abstract them as virtual nodes

22. ATRI: Adaptive Triangular Deployment Algorithm • Snapshots of the execution for the environment with obstacles

23. ATRI: Adaptive Triangular Deployment Algorithm • 2. Non-uniform deployment • The density of nodes can be adjusted adaptively to different requirements of tasks • Strategy: set a shorter sensing range in important area

24. ATRI: Adaptive Triangular Deployment Algorithm • Snapshots of the execution for non-uniform deployment

25. Performance Evaluation • Measurement of performance • Deployment quality: total coverage area • Moving energy consumption: moving distance • Comparison with VEC in [G. Wang, G. Cao, and T. La Porta, “Movement-Assisted Sensor Deployment,” INFOCOM, 2004]

26. Performance Evaluation Total coverage area gets larger

27. Performance Evaluation Average moving distance gets smaller

28. Conclusion • The optimal node layout for maximum coverage • Equilateral triangulation (the length of each side equals ) • Basic Triangular Algorithm • Divides the transmission circle into six sectors and adjust the relative distance between neighbors • For reducing node oscillation • Distance threshold strategy • Movement state diagram strategy • Adaptive Triangular algorithm • Avoiding obstacles and boundaries • Abstract them as virtual nodes • Non-uniform deployment • Set a shorter sensing range in important area

29. Discussion • The lack of mentions about • Relation between communication range and sensing range • Definition of some notations • The meaning of some equations • Termination condition • No consideration about • Communication overhead • Deployment time • Impact of threshold • Synchronization and collision

30. Thank you

31. References • [1] I. Akyildiz, W. Su, Y. Sankarasubramaniam, and E. Cayirci, “Wireless Sensor Networks: A Survey,” Computer Networks, 2002. • [2] G. Wang, G. Cao, and T. La Porta, “Movement-Assisted Sensor Deployment,” INFOCOM, 2004

32. Appendix A: Delaunay Triangulation • What is Delaunay Triangulation?

33. Appendix B: Minimum average moving distance • Minimum average moving distance • Proof