1. EASE: An Energy-Efficient In-Network Storage Scheme for Object Tracking in Sensor Networks Jianliang Xu Department of Computer Science Hong Kong Baptist University Hong Kong Xueyan Tang School of Computer Engineering Nanyang Technological University Singapore Wang-Chien Lee Department of Computer Science & Engineering Penn State University University Park, PA IEEE SECON 2005

2. Outline Introduction Related works EASE: Energy-Conserving Approximate Storage Performance analysis and simulations Conclusions

3. Introduction Energy efficiency is one of the most critical issues in the design of wireless sensor networks Many sensor applications for object tracking can tolerate a certain degree of imprecision in location data of tracked objects

4. Related works Research efforts towards energy conserving object tracking sensor networks [9]-[12] Reduce the number of sensing nodes activated for tracking an object Reduce the location updating traffic in providing accurate answers to location queries

5. Introduction Reduce the location updating traffic Centralized / designated storage for data collection and query answering

6. Introduction Local storage (LS) Sensing data is stored in local memory of sensors. sink

7. Centralized Storage (CS) Data is stored in data-centric nodes. sink Query: Car 1 (Car 1) (Car 3) (Car 2) Introduction Detecting Node

8. Introduction - Motivation- We take a different approach to improve energy efficiency by exploiting the tradeoff between data quality and energy conservation

9. EASE - System Model - We consider a sensor network consisting of a large number of stationary sensor nodes deployed in some operational area Each sensor node is aware of its own location

10. EASE Approximation radius Species a geographical area in which the low precision location data is considered to be valid sink

11. EASE - Approximation radius - Optimal Setting of approximation radius

12. EASE EASE innovatively maintains two versions of object location data in the network Local storage node Stores high-precision data Data centric node Stores low-precision data Be an index server that keeps track of the local storage nodes

13. EASE - Detect - Approximation Area Defined by a circle with o as the center and r as the radius Static Sensor Node Data Centric Node Querying Node Detecting Node Car 1 Car 2 Car 1 Local storage Node r

14. r EASE - Report - A location update due to object movement will not be sent to the designated storage node if the object remains within the approximation radius Static Sensor Node Data Centric Node Querying Node Detecting Node Car 1 Car 2 Car 1 Local storage Node Update Report

15. EASE - Report - Static Sensor Node Data Centric Node Querying Node Detecting Node Car 1 Car 2 Car 1 Local storage Node Query a: Response Query Forwarding Update Report Response r

16. EASE - Local Update - Static Sensor Node Data Centric Node Querying Node Detecting Node Car 1 Local storage Node Update Report r

17. EASE - Remote Update (Case 1)- Static Sensor Node Data Centric Node Querying Node Detecting Node Car 1 Local storage Node Geocast Message for Invalidation Update Report r’ r

18. r EASE - Remote Update (Case 1)- Static Sensor Node Data Centric Node Querying Node Detecting Node Car 1 Local storage Node Geocast Message for Invalidation Update Report r’

19. Summarizations Addressing energy efficiency by exploiting the trade-off between data quality and energy conservation Efficiently answer precision constrained approximate location queries

20. Performance Analysis

21. Message complexity of EASE Data Centric Node Detected Node Local storage Node (1) (2) (1)(2) The cost required for reporting data The cost required for update

22. Performance Analysis Cost of remote update (C ur ) Cost of local update (C ul ) The number of sensor nodes in the approximation area

23. Performance Analysis P qf (r)C qn Remote update Local update

24. Optimal Approximation Setting We assume a 2-dimensional random walk model It moves a distance of d along an arbitrary direction at each step Each step takes a duration of l The average time an object takes to move out of a circle of radius r

25. Optimal Approximation Setting

26. Simulations - Parameter -

27. Simulations - Mobility profile setting -

28. Message Complexity vs. Precision Requirement

29. Breakdown of Message Complexity

30. Message Complexity vs. Query Rate

31. Energy Consumption vs. Precision Requirement

32. Energy Consumption vs. Query Rate

33. Conclusions This paper presents a study on precision- constrained approximate location queries for object tracking sensor networks. An energy-efficient in-network storage scheme called EASE has been developed to efficiently Reduce the network traffic Conserve the energy of sensor nodes Improve the network lifetime

34. Thank You !!!