Sensor Network Assisted Collaboration for Pursuit-Evasion Problem

Abstract

The pursuit-evasion problem represents many important real-world applications. In this paper, we present a new adaptive pursuer collaboration method with the support of wireless sensor networks. We focus on two different evader motions, random movement and active escaping, and develop a moving pattern recognition algorithm based on information obtained by the sensor network. We compare two pursuing strategies, greedy vs. collaboration, and derive their expected capturing times for the two evader moving patterns. By recognizing the evader moving patterns and selecting appropriate pursuing strategies, the new collaboration method is more adaptive, robust, and reduces the capturing time significantly in situations where the number of pursuers is small and the operating field is large