基于粒子滤波的多传感器交互式多模型多机动目标跟踪

摘要

针对单传感器交互式多模型联合概率数据关联滤波算法(Interacting Multiple Model Joint Probabilistic Data Association Filtering,IMMJPDAF)在非线性情况下跟踪精度低,且对于非高斯问题不适用的情况,文中提出一种基于粒子滤波的多传感器交互式多模型多机动目标跟踪算法(Interacting Multiple Model Joint Probabilistic Data Association Particle Filtering,IMM-JPDA-PF).将IMM,JPDA和PF相结合,给出了两个传感器情况下的IMM-JPDA-PF算法,并且IMM-JPDA-PF()算法能够很容易地扩展到任意多个传感器的情况,在非线性非高斯条件下实现了杂波环境中利用多传感器对多机动目标的有效跟踪.仿真结果表明,多传感器IMM-JPDA-PF算法比单传感器的IMM-JPDA-PF算法具有更高的多机动目标跟踪精度.%The interacting multiple model joint probabilistic data association filtering(IMMJPDAF) algorithm of single sensor has low tracking accuracy in nonlinear cases, and it is not appropriate for non-Gaussion situation. To solve the problems, an interacting multiple model target tracking algorithm of multiple sensor multiple maneuvering targets based on particle filter (IMM-JPDA-PF) is proposed. Combined with IMM, JPDA and PF, the IMM-JPDA-PF algorithm of two sensors is given, which can be easily expanded to the case of any number of sensor. The multiple sensor IMM-JPDA-PF algorithm can track efficiently multiple maneuvering target with multiple sensor under the condition of nonlinear non-Gaussion in clutter environment. Simulation results demonstrate that the multiple sensor IMM-JPDA-PF algorithm has higher tracking accuracy than that of single sensor IMM-JPDA-PF algorithm.