A Parallel Computing Algorithm for Moving Targets Tracking in Wireless Sensor Networks

摘要

In this paper, we address the problem of moving target feature extraction and collaborative tracking in wireless sensor networks (WSN), and present a parallel computing algorithm for moving human collaborative tracking. At first, WSN optimization deployment: divide the monitors in WSN into the two types: Behavior Recognition Monitor (BRM) and Collaborative Tracking Monitor (CTM), and settle all the monitors utilizing FCM algorithm into many groups. Secondly, parallel detection and behavior recognition to get the collaborative tracking target. Finally, a multi-points feature extraction scheme for WSN monitors to track the suspicious target collaboratively. We also compare our algorithm with three existing solutions, the statistics result shows that our scheme has a better detection accuracy and tracking performance.