A Semi-Localized Algorithm for Cluster Head Selection for Target Tracking in Grid Wireless Sensor Networks.

Wireless sensor networks are widely used in many application domains. One of the well-known applications of such networks is tracking of moving objects. In this paper we propose a novel algorithm in which head and members of next clusters are chosen and woken up dynamically by the current active cluster head in a grid network of wireless sensors. This dynamic algorithm is designed for target tracking applications. The proposed solution is a distributed clustering scheme, in which, head and cluster members of the next cluster are selected cooperatively. Based on the distributed solution, a semi-localized clustering scheme is also presented. Moreover, a distributed target movement history recording algorithm is proposed to reduce target misses. Simulation results show that by using the proposed method, there are significant reductions in the average number of received messages, averaged number of sent messages, and average number of samplings of the sensing subsystem. In addition, a meaningful reduction in target miss rate is achieved by using distributed target movement recording algorithm. [ABSTRACT FROM AUTHOR]