Energy Efficiency Research On Target Tracking In Binary-Sensor Network

Recent developments in theories and methodologies of Wireless Sensor Networks (WSNs) have made a wide of applications possible, among which target tracking is one of the typical ones. Featured in eases of deployment, installation and extension, WSNs facilitate the researches in target tracking and make it feasible for tracking in complex environment. However, resource constraints in hardware and energy induce new problems for target tracking with WSNs. To address them, lots of research has been attracted and many algorithms have been proposed. In this thesis, in view of the ease of practical realization of binary sensors, we do a lot of studies in target tracking with WSNs with binary sensors, with accounting for energy constraints.The thesis first introduces the principles and basic processes of target tracking with WSNs. Then, it analyzes and proposes several algorithms for target localization, sensor activation and target prediction issues, respectively. Followed by the analysis of target localization, a precise localization algorithm with compensation for the existing centroid localization algorithm with selective sensor activation mechanism is presented, which is proved to improve efficiency in target tracking. To improve the efficiency of sensor activation for target tracking with WSNs, we propose an adaptive sensor activation algorithm, which dynamically adjusts the activation ranges according to targeting movements instead of using a fixed one. In this case, manually configuring the activation ranges is not necessary any more. Besides, the tracking system can work in various environments and the energy efficiency is also improved. For target prediction, we advance the interacting multi-model (IMM) algorithm into road-based maneuvering target prediction, and propose a road-based sensor activation (RBSA) algorithm, which adopts different prediction models for different road types. It is shown that RBSA can achieve similar tracking performances with existing complex multi-model-based algorithms, but lower requirement with computations and prediction precision. In addition, tracking of targets with unpredictable maneuver is also considered in this thesis. A distributed sensor activation algorithm (DSA~2) based on duty-cycle activation algorithm is designed to improve the tracking successfulness and also the energy efficiency.Finally, the thesis concludes the four algorithms for the improvements of tracking efficiency. And some future research directions are highlighted.