低检测概率条件下的多传感器机动多目标跟踪方法研究  

A Tracking Method of Multi-sensor to Track the Multiple Targets Under the Condition of Low Detection Probability

作　　者：倪龙强[1,2] 高社生[1] 薛丽[1] 

机构地区：[1]西北工业大学自动化学院,陕西西安710072 [2]西北机电工程研究所,陕西咸阳712099

出　　处：《兵工学报》2013年第1期87-92,共6页Acta Armamentarii

基　　金：国家自然科学基金项目(61174193);航空科学基金项目(20080816004)

摘　　要：为解决低检测概率条件下的多传感器非线性、机动、多目标检测、数据关联及滤波问题,首先对目标数量进行随机过程建模,其次应用模型参数以及目标数量对目标状态进行了增广,最后应用多模型粒子滤波器(MMPF)对多传感器在低检测概率条件下的机动多目标跟踪进行了仿真。仿真结果表明:基于MMPF的低检测概率目标跟踪方法能够有效检测目标数量,同时对机动多目标具有良好的跟踪性能。A new algorithm was presented to deal with the multi-target tracking problem.Firstly the target number in an interest region was modeled as a stochastic process;secondly the state vector was augmented with target number;and finally the state estimation was carried using the multi-model particle filter（MMPF）,and the numerical simulation was proposed to identify the efficiency of this method in multi-sensor/multi-target tracking application.The simulation results show that the improved method can be applied to track the maneuvering targets effectively by using the non-linear dynamic model.

关 键 词：兵器科学与技术 多传感器融合 目标跟踪 粒子滤波 数据关联 

分 类 号：TN593[电子电信]