Research and Improvement on Tracking Algorithm in Wireless Sensor Networks
|School||Nanjing University of Aeronautics and Astronautics|
|Course||Applied Computer Technology|
|Keywords||Wireless sensor networks Tracking Particle filter Prediction mechanism|
Target tracking based on sensor networks have broad application prospects, such as wildlife, medical research and military intelligence gathering , earthquake relief and other areas are of great significance. Goal of this study is to make these applications we can accurately grasp the real-time location tracking target , as well as in the context of limited energy of nodes extending sensor network lifetime . The main work is as follows : ( a ) respectively, based on the binary sensor network model and the network model based on precise positioning sensors under both architectures tracking algorithm analysis and research. ( 2 ) focus on two different architectures Typical tracking algorithm OccamTrack and Cluster-based predictive tracking algorithm mechanisms are inadequate to start from the structural characteristics of sensor networks , target tracking accuracy, energy consumption point of view of the shortcomings of the algorithm and proposed improved method feasible . ( 3 ) the use of the idea of ??particle filtering for binary sensor network model characteristics and geometry of OccamTrack algorithm is studied and improved, presents a new particle filter based geometry processing algorithms. ( 4 ) for precise positioning based clustering model WSN target tracking algorithm, using Bayesian estimation improved distributed cluster head node prediction mechanism and made a recovery track failure tracking mechanism . ( 5 ) to build simulation program simulation platform and writing , respectively, of the two algorithms to improve the simulation results and detailed comparative analysis of the results seen from the simulation , two improved algorithm in target tracking accuracy, the angle of the energy consumption better than the original algorithm .