Distributed Positioning and Tracking in Cluster-based Wireless Sensor Networks

In this paper, distributed positioning and tracking schemes for cluster-based wireless sensor networks (WSNs) are presented. We focus on the cases with overlapping cluster regions, which are particularly useful in applications involving inter-cluster communications, time synchronization, and target tracking. In overlapping clustering, the duplicated correlated data between adjacent clusters may, however, impair the performance in information fusion. To tackle the problem, a geometric formation with regularly distributed overlapping clusters is used for situation analysis. Each cluster is composed of one cluster head (CH) and a number of normal sensors. Different operating modes are defined for the CH and normal sensors. To resolve the problem of correlated data, Cholesky decomposition is adopted to decorrelate the measurement noises. In addition, extended information filtering (EIF) is modified for distributed target estimation. The simulations and results show that the proposed distributed schemes with the designated mode settings attain good efficiency and accuracy in target tracking.