交互多模型的Rao-Blackwellized粒子滤波算法在多目标跟踪中的应用  

Multiple Target Track Based on Interacting Multiple Model Rao-Blackwellized Particle Filter Algorithm

作　　者：张万里[1] 何金刚[1] 

机构地区：[1]中国空空导弹研究院,河南洛阳471009

出　　处：《航空兵器》2014年第4期3-7,共5页Aero Weaponry

基　　金：航空科学基金资助项目(20110112006);重点项目资助(20110112001)

摘　　要：在雷达信号处理领域,运动目标跟踪问题一直是研究的重点,而非线性滤波问题则是难点所在。本文以Rao-Blackwellized粒子滤波(RBPF)算法为基础,针对其应用于雷达目标跟踪时产生的算法收敛情况较差问题进行研究。为提高算法的收敛性能及滤波精度,对RBPF算法中线性部分采用交互多模型的方法进行状态预测和校正,利用数字仿真对更改后的算法进行验证,结果证明改进后的RBPF算法在各个方面均有较大程度改善。In radar signal processing field , moving target tracking is always the focus , it involved the nonlinear filter problem which is the difficult point .In this paper , the problem that when the radar tracks target, the convergence of Rao-Blackwellized particle filter algorithm becomes worse is studied .To im-prove the convergence and the precision of the filter algorithm , interacting multiple model method in the RBPF algorithm is used to forecast and correct status , which is validated by numerical simulation .The simulation result shows that the improved RBPF algorithm is enhanced in every aspect .

关 键 词：雷达目标跟踪 制导信息估计 粒子滤波 RBPF算法 交互多模型 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]