摘要
利用有限集统计理论对声学W SN下的多目标跟踪进行了研究。首先利用F ISST理论对声学W SN下的多目标跟踪进行建模,然后利用粒子滤波对随机集最优贝叶斯滤波进行近似,提出了一种新的多目标跟踪方法。在粒子滤波执行过程中,为了有效地平衡计算量和跟踪精度,根据目标数目自适应地调整每一时刻的粒子数目。仿真结果表明在监视区域同时出现目标较少的情况下,算法能够及时发现目标的出现和消失,正确估计目标状态和数目。
This paper studies multi-target tracking in WSN using finite set statistic （FISST） theory. Firstly,the multi-target tracking in aoustic WSN is modeled by using FISST theory, and then by using particle filter to approximate optimal Bayes filter, an algorithm of tracking a variable number of targets in acoustic WSN is proposed. In the process of particle filter, to balance the computation complexity and tracking precision, the number of particles is adjusted automaticly according to the number of targets. The results of simulations show that the algorithm has good perform-ance in estimating the number and states of targets when there are not many targets in the surveillance area.
出处
《火力与指挥控制》
CSCD
北大核心
2012年第7期24-27,共4页
Fire Control & Command Control
基金
国家863计划项目(2007AA01Z309)
湖北省自然科学基金资助项目(2009CDB301)
关键词
多目标跟踪
有限集统计理论
粒子滤波
概率假设密度
Multi-target tracking, finite sets statistic theory, particle filter, probability hypothesis density
