Trapping Mobile Targets in Wireless Sensor Networks: An Energy-Efficient Perspective

Abstract

Mobile target detection is a significant application in wireless sensor networks (WSNs). In fact, it is rather expensive to require every part of the region of interest (RoI) to be covered in a large-scale WSN for target detection. Trap coverage has been proposed to trade off between sensing performance and the cost of sensor deployments. It restricts the farthest distance that a target can move without being detected rather than providing full coverage to the region. However, the results cannot be directly applied in a real WSN since the detection pattern of a sensor in practical scenarios follows a probabilistic sensing model. Moreover, the trap coverage model does not consider the various moving speeds of targets, which is important for trapping targets. To extend the concept of mobile target trapping into a real large-scale WSN, we analyze the detection probability of a mobile target in the sensor network theoretically and define probabilistic trap coverage in this paper, which restricts the farthest displacement of a mobile target with a detection probability less than the threshold. We develop the theory of circle graph, which can be generally applied in the area of intrusion detection such as trap coverage and barrier coverage. We further study the practical issue of how to schedule sensors to maximize the lifetime of a network while guaranteeing probabilistic trap coverage. A localized protocol is proposed to solve the problem, and the performance of the protocol is theoretically analyzed. The lower bound of lifetime acquired by the protocol is nearly half the optimum lifetime. To evaluate our design, we perform extensive simulations to compare our algorithm with the state-of-the-art solution and demonstrate the superiority of our algorithm.