基于Fisher信息距离的被动传感器目标协同跟踪方法  

Target collaboration tracking algorithm for passive sensor based on Fisher information distance

作　　者：王勇[1,2] 

机构地区：[1]中国矿业大学信息与电气工程学院,江苏徐州221116 [2]徐州工业职业技术学院,江苏徐州221140

出　　处：《传感器与微系统》2014年第7期119-122,共4页Transducer and Microsystem Technologies

摘　　要：针对多被动传感器动态跟踪问题,提出了一种基于Fisher信息距离被动传感器目标协同跟踪方法。该算法在进行传感器选择时,依据信息几何理论,以流形中的Fisher信息距离来衡量先验概率密度函数和后验概率密度函数之间的距离,继而以此距离为依据选择传感器进行目标跟踪。仿真实验表明:所提算法能够在动态环境中自适应选择传感器资源,有效提高目标的跟踪精度,实现多被动传感器协同跟踪。Aiming at dynamic tracking problem of multiple passive sensors,a passive sensor target collaboration tracking algorithm based on Fisher information distance is proposed.According to the theory of information geometry,the distance between the posterior probability density and the prior probability density function is defined by Fisher information distance in manifold,which is used as the basis for sensors selecting to carry out targe tracking.Simulation results show that the proposed method can adaptively select resource of sensor in dynamic condition,effectively improve tracking precision and realize multi-passive sensor collaboration tracking.

关 键 词：目标协同跟踪 信息几何 Fisher信息距离 被动传感器 

分 类 号：TN953[电子电信—信号与信息处理]