Study On Target Tracking In Wireless Sensor Networks Algorithm Based On Distributed Particle Filtering

As an emerging IT hotspot technology, Wireless sensor networks has broad applicationprospects in many areas of military and civilian. Since sensor nodes in sensor networks withlimited processing power and limited energy to carry, Therefore, the design of simple andefficient target tracking algorithm for sensor networks to extend the life cycle and to enhancethe performance of sensor networks has important significance.The complexity of the target tracking and Wireless sensor network node insufficientcapacity to form contradictory, distributed is a powerful way to resolve the contradiction.According to the movement of the target position, the sensor nodes of wireless sensornetworks dynamically organized clusters, and established the distributed tracking structure. Toimprove the performance of target tracking, and reduce energy costs of wireless sensornetworks, distributed information particle filter algorithm and distributed parallel particlefilter algorithm had been proposed. Use of information extended Kalman filter combines thelatest observations, produce the particle filter proposal distribution, the specific steps torealize distributed information particle filter which is based on dynamic clustering structurehad been introduced in detail. To complete parallel particle filtering process,distributedparallel particle filter algorithm uses parallel particle filter to divide particle collection intomultiple small subset of particles, and then assigned to each sub-node in the cluster.In order to realize distributed information particle filter algorithm and distributed parallelparticle filter algorithm, and effectively reduce energy consumption and other issues inwireless sensor networks for target tracking, target tracking dynamic clustering simulationenvironment in wireless sensor networks had been establish, and then simulation analysis ofthe two algorithms had been carried out. Simulation results show that, the two algorithms ofthe node has greatly improved energy consumption. Distributed parallel particle filteralgorithm uses parallel processing approach to calculate the valuation which not onlyimproves the issue of energy consumption, but also improves the tracking accuracy at thesame time.