Research On Distributed Kalman Prediction Filter-Based Tracking Algorithm

As the electronic technology and system integration technology development, some miniature embedded sensors which have the abilities to perception, memory, computing and communication are begun to appear. Its applications, especially the application of tracking the moving target, are attended greatly by the many people. The technology of tracking the moving target can be widely used in military, civilian and family health and so on, such as the enemy reconnaissance, navigation, children’s education, etc.In order to solving the problem about tracking the moving target,this paper introduced the kalman filtering algorithm, and provided a distributed target tracking algorithm with dynamic clustering structure. In this algorithm, the wireless sensor network nodes used dormancy/activate mechanism the neighbors distance guide probability density of scheduling mechanism to select nodes which would be organized into a cluster. These nodes in the cluster would collect information to predict and estimate the position of moving target according to kalman filtering algorithm. We can obtain the trajectory of moving target finally through the nodes constant circularly organized dynamic cluster and constant updated predicted values. This paper established the tracking simulation environment about wireless sensor network and compared the accuracy of tracking in different conditions which different nodes coverage and different maximal detection radius. The simulation results show that the tracking accuracy of the algorithm can better the application requirements. And this paper conducts some related verification experiments about the algorithm. The experiments results show the feasibility of the algorithm.