无线传感器网络中基于信息收益的多目标跟踪

A Method of Multi-Target Tracking Based on Information Utility in Wireless Sensor Networks

目标跟踪是无线传感器网络应用的一个很重要的研究领域,节点有限的通信能力、处理能力、存储能力限制了传统跟踪算法的使用。为了提高网络的信息感知能力和降低能量消耗,提出基于信息收益的加权质心多目标跟踪算法,算法通过信息收益函数选取并唤醒节点对目标感知,通过传感器属性信息解模糊,并利用加权质心算法实现了多目标跟踪。仿真结果表明,与Bayes滤波协同多目标跟踪算法相比,算法虽精度略低,但时间复杂度低、失跟率低、实时性高,总体性能优于前者。

Tracking is a significant area in the application of wireless sensor networks. However, the limitation of nodes in the aspects of communication, processing and storage have confined the use of traditional methods in tracking. In order to improve the ability of information perception of networks and reduce energy consumption, a method of weighted centroid multi -target tracking based on information utility was proposed. In this method, the nodes were selected and wakened to sense targets by using information utility function, illegibility was dealt with through attribute information of sensors, and ultimately, by using weighted centroid method, muhi - target tracking was realized. The simulation indicates that this method has lower computational complexity and missing target ratio, and higher real - time capability, except for slightly lower tracking precision, compared with the Bayesian filer method.