Decentralized Sensor-Coordination Optimization for Mobile Multi-Target Tracking in Wireless Sensor Networks

To minimize energy consumption in the Wireless Sensor Networks (WSNs), we propose a decentralized sensor coordination optimization scheme for Mobile Multi-Target Tracking (MMTT) in WSNs. Our scheme partitions the available sensor-nodes into clusters using the maximum-entropy based clustering criteria. For each tracked target, a number of neighboring clusters are activated based on their Hausdorff distance to the tracked targets.We propose the corresponding target-position estimation scheme using the particle Probability Hypothesis Density (PHD) filtering algorithm. Furthermore, our sensor coordination scheme dynamically selects the cluster members to jointly optimize the sensing accuracy and the energy efficiency by using our proposed Decentralized Particle Swarm Optimization (DPSO) based algorithm. The conducted performance simulations evaluate our proposed sensor-coordination schemes, which show the optimal performance in terms of energy efficiency of the WSN while maintaining a high target-position estimate accuracy.