Dynamic clustering in wireless sensor network for target tracking based on the fisher information of modified Kalman filter

Abstract

In order to reduce the whole network energy consumption of wireless sensor network and select the most suitable nodes to participate in target tracking, a dynamic clustering method is proposed using an improved Kalman filter which based on Fisher information matrix in target tracking. With the basis of using of information criteria as the selection criterion of nodes in target tracking, the node residual energy is joined as a selection criteria about voting cluster-head nodes and carrying on the dynamic clustering in this network. At the same time, it can describe the restructuring of dynamic cluster and switch of the cluster head. Cluster head receive the estimations of target from their cluster member nodes, then, it uses the modified Kalman filter (KF) based on Fisher information matrix for filtering in target tracking. Simulation of this method comparing with clustering controlling of activated radius and non-clustering information matrix filter, results show that the proposed method can effectively control the number of the trace nodes, at the same time, improving the tracking precision.