高转弯频率下机动目标跟踪算法研究  

Research on maneuvering target tracking algorithm under high turn frequency circumstances

作　　者：张杰[1] 张建伟[1] 

机构地区：[1]四川大学计算机学院,成都610064

出　　处：《计算机工程与应用》2011年第1期217-219,244,共4页Computer Engineering and Applications

基　　金：国家高技术研究发展计划(863)重点项目No.2006AA12A104~~

摘　　要：UKF_IMM算法对任何非线性高斯系统都有较好的跟踪性能,但在用于强非线性、非高斯系统时将产生极大的误差。PF_IMM算法适用于任何非线性非高斯系统,但它因计算量很大而导致实时性太差。在现有的基于PF或UKF的IMM滤波算法中,PF或UKF通常被用于每一个模型,基于此,提出了IUS_IMM算法,即在转弯模型中使用标准粒子滤波器(SPF),在其他模型中用UKF滤波器,仿真结果表明,对于转弯频率高的机动目标I,US_IMM的性能要好于UKF_IMM算法。The algorithm of UKF_IMM has better tracking performance for arbitrary nonlinear systems under the Gaussian conditions.But serious errors can arise when it is used in nonlinear nongaussian systems.PF_IMM is effective in nonlinearity and nongaussianity,however,it has a large computation.In existing literature on combining IMM and PF or UKF techniques to tackle difficult target maneuvers,a PF or UKF is usually used in every model.In comparison,IUS_IMM which uses a PF in the 2DTR model and UKF in the remaining models is put forward.Simulation results show that IUS_IMM has superior performance than UKF_IMM when maneuvering target is under high turn frequency circumstances.

关 键 词：无迹卡尔曼滤波器 粒子滤波器 机动目标跟踪 交互多模型 

分 类 号：TP39[自动化与计算机技术—计算机应用技术] TN95[自动化与计算机技术—计算机科学与技术]