基于无味粒子滤波和交互多模型算法的多机动目标跟踪  

Multi-target tracking based on unscented particle filter and interacting multiple models algorithm

作　　者：何祖军[1] 尚明玲[1] 

机构地区：[1]江苏科技大学电子信息学院,江苏镇江212003

出　　处：《江苏科技大学学报（自然科学版）》2008年第6期48-52,共5页Journal of Jiangsu University of Science and Technology:Natural Science Edition

基　　金：江苏省高技术研究资助项目(BG2006022)

摘　　要：闪烁噪声是一种非高斯噪声.为了提高闪烁噪声下多机动目标跟踪的精度,在交互多模型IMM(Interacting Multiple Models)算法的基础上将非线性非高斯系统滤波算法——粒子滤波与IMM算法相结合,采用无味粒子滤波UPF(Unscented Particle Filter)代替IMM算法中各模型的卡尔曼滤波,提出了一种UPF_IMM算法,并应用该算法代替传统IMM_JPDA数据关联方法中的IMM部分,解决了闪烁噪声环境下的多目标跟踪问题,实验结果表明该算法可以明显地提高跟踪精度.Glint noise is a kind of non-Gaussian noise. In order to improve tracking precision of multi-target under glint noise, UPF_IMM （ Unscented Particle Filter Interacting Multiple Models） algorithm is proposed based on IMM algorithm in this paper. Also, nonlinear and non-Gaussian fiher--Particle Filter（PF） is integrated with IMM algorithm. Unscented PF （UPF） is applied to take the place of Kalman filter in IMM algorithm. Furthermore, UPF_IMM is used to take the place of IMM in traditional data association algorithm IMM _JPDA; and then multi-target tracking under glint noise is resolved. Simulation results indicate the proposed algorithm can obviously enhance tracking precision.

关 键 词：多目标跟踪 粒子滤波器 IMM算法 闪烁噪声 

分 类 号：TN953[电子电信—信号与信息处理]