Target Detection And Tracking In Wireless Sensor Networks

With rapid advances in Microelectromechanical Systems (MEMS), wireless communication, computing and sensors technologies, Wireless Sensor Networks (WSNs), integrated with sensing, processing and communication, have become the research frontier in recent years. Because of their unique characteristics, WSNs have been widely used in many applications, such as military defense, environment monitoring, intelligent house and medical attendance. Therefore, a WSN has been considered as one of the21most important technologies for the21st century and one of the10emerging technologies that will change the world.Target detection and tracking is one of the fundamental problems for WSNs and plays an important role in the safety field. In general, a WSN is large-scale and resource limited. This thesis focus on target detection and tracking in WSNs and mainly studies sensor deployment, sensor management and scheduling, and real-time data reporting. The contributions of this thesis are summarized as follows:1. For WSNs with stationary sensors, it is difficult to form barrier coverage after initial random deployment. To solve this problem, we leverage mobile sensors to fill in the gaps between stationary sensors to form barrier coverage. We introduce the notion of weighted barrier graph and prove that the minimum cost of mobile sensors needed is equivalent to the shortest path on the weighted barrier graph. In addition, we formulate the minimum cost mobile sensor movement problem as a minimum cost bipartite assignment problem. Based on the analysis, we propose a directed barrier coverage algorithm to form barrier coverage using the minimum cost of mobile sensors. Theoretical analysis and simulation results validate the effectiveness of the algorithm.2. Considering that sensor nodes usually have location errors in real deployment, we study the influence of location errors on barrier coverage formation. We analyze the influence of loca-tion error on the true minimum number of mobile nodes needed, and propose a progressive mobile node deployment algorithm that schedules mobile nodes to expected locations to fill in gaps between stationary sensor nodes. In addition, we introduce a fault tolerant weighted barrier graph to model the barrier coverage formation problem with location errors and find the minimum cost of mobile nodes needed to guarantee the formation of a barrier. Exten-sive simulation results validate the influence of location errors and also the correctness of analysis.3. For target tracking in large-scale WSNs, we propose a cluster-based target tracking algo-rithm, which leverages the cluster structure to wake up and sleep sensor nodes, and routes the estimated target locations through the backbone composed of cluster heads. Experimen-tal results show that the cluster-based target tracking algorithms can energy-efficiently track a target and report target locations in real-time manner.4. Cluster-based target tracking algorithms suffer from boundary problem. To solve the prob-lem, we propose a hybrid cluster-based target tracking algorithm which integrates dynamic clustering into static cluster structure. As the target moves in the network, dynamic clus-ters and static clusters alternatively track it. Experimental results show that the algorithm can efficiently solve the boundary problem, and outperforms other typical target tracking algorithms.5. In WSNs, few of existing target tracking algorithms have been implemented and evaluated on real systems. To efficiently evaluate our proposed algorithms, we design and implement a real target tracking system, which consists of36sensor nodes, a sink node and a base station. Extensive localization and tracking experiments validate the effectiveness of the target tracking system.