【作者】 颜振亚； 郑宝玉； 崔景伍；

【Author】 YAN Zhen-ya ZHENG Bao-yu,CUI Jing-wu Institute of Signal & Information Processing,Nanjing University of Posts and Telecommunication,Nanjing 210003

【机构】 南京邮电大学信号与信息处理研究所；

【摘要】 本文讨论如何利用高斯和粒子滤波(Gaussian sum particle filter)来进行协作目标跟踪。协作目标跟踪是通过节点选择来完成的,跟踪的结果在选择的节点之间传输。为了节省通信资源,本文利用高斯和粒子滤波来进行目标跟踪,此时,仅仅需要传榆均值和方差就能表示跟踪结果。同时,本文引入了均值偏移(mean shift)过程来提高高斯和粒子滤波的性能,并基于高斯和粒子滤波提出了一种节点选择算法。最后,通过仿真来验证算法的有效性。更多还原

【Abstract】 In this paper, a novel framework of distributed target tracking via Gaussian sum particle filter is proposed. The collaborative tracking is implemented through sensor selection, and the result of tracking is proposed among sensor nodes. In order to save the communication resources, we employ the Gaussian sum particle filter to perform the target tracking, in which pnly the mean and covariance of mixands need to be communicated. Also the mean shift to improve the tracking performance is introduced and a sensor selection criterion based on Gaussian sum particle filter is presented. Simulation results show that our method works well.更多还原