Prediction-based Proactive and Reactive Cluster Managements for Target Tracking in Binary Sensor Networks

A layered scheme for target tracking in binary sensor networks is proposed in this paper to improve the tracking precision with minimum resource allocation. In the physical layer, binary sensors are employed, reducing not only hardware configuration overhead but also communication cost. By adopting binary variational filtering algorithm, considerable accurate tracking is ensured even in case of non-linear and non-Gaussian environment. While communication between Cluster Heads (CHs) decreases dramatically compared to particle filtering algorithm. For different requirements, two kinds of cluster management strategies are proposed: Proactive cluster management (ProCM) and Reactive Cluster Management (Re-CM). The former strategy aims at minimizing hardware expenditure and total energy consumption of the binary sensor network (BSN). Whereas the main concern of the latter is to maximize BSN lifetime by evenly distributing energy expenditure over the whole BSN. The effectiveness of the proposed schemes are evaluated and compared in terms of tracking accuracy, energy consumption and execution time.