适于无源阵列跟踪机动目标的IMM-PF算法  

IMM-PF Algorithm for Passive Array Tracking of Maneuvering Target

作　　者：郭艺夺[1] 童宁宁[1] 王凯[1] 

机构地区：[1]空军工程大学导弹学院,陕西三原713800

出　　处：《电光与控制》2009年第1期39-41,46,共4页Electronics Optics & Control

基　　金：国家"八六三"计划资助项目(2006AA701307)

摘　　要：针对无源阵列对机动目标跟踪效果较差的问题,在融合交互式多模型和粒子滤波方法的基础上,提出了一种基于粒子滤波的交互多模型(IMM-PF)算法。该算法采用多模型结构以跟踪目标的任意机动;各模型采用粒子滤波算法,以处理非线性、非高斯问题。各模型中相对固定数目的粒子群经过相互交互、粒子滤波后再进行重抽样以减少滤波退化现象。与典型的交互式多模型算法(IMM-KF)进行了比较,计算机仿真结果证实了新算法的正确性和有效性。Passive array is generally not so effective in tracking maneuvering target. To solve the problem, an Interacting Multi-Model Particle Filter（IMM-PF） algorithm is presented based on Interacting Multiple Model（IMM） and particle filter. The algorithm uses multiple models to track arbitrary maneuvering of the target. Each model uses particle filter to deal with the nonlinear/non-Gaussian problems. After interaction and particle filtering, the particles with relatively fixed number in each model are resampled to reduce the degeneracy of filtering In the simulations, the algorithm is compared with the general interacting multi- model, and the results demonstrate the correctness and efficiency of the new filtering method.

关 键 词：机动目标跟踪 交互式多模型 粒子滤波 非线性/非高斯问题 重抽样 

分 类 号：V271.4[航空宇航科学与技术—飞行器设计] TN911.7[电子电信—通信与信息系统]