低检测概率条件下的多传感器机动多目标跟踪方法研究

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.