基于多模型粒子滤波的机动多目标跟踪算法

Maneuvering Multiple Target Tracking Algorithm Based on Multiple Model Particle Filter

针对密集杂波环境下机动多目标跟踪中系统强非线性以及运动模式切换对于滤波精度的不利影响,提出了一种基于多模型粒子滤波的机动多目标跟踪算法。新算法实现了多模型粒子滤波和广义概率数据关联算法的有机结合。通过在粒子状态采样过程中引入模型信息改善了交互式多模型和粒子滤波结合中导致的计算量膨胀问题,并利用广义概率数据关联算法实现回波的有效确认和回波信息的充分利用。给出了应用该方法的具体步骤,最后,理论分析和仿真实验证明该算法的有效性。

To eliminate the adverse impact on filter precision which was brought about by the maneuvering multi-target tracking system’s strong nonlinear and motion model switching in clutters environment,a novel maneuvering multi-target tracking algorithm based on multiple model particle filter was presented. The dynamic combination of multiple model particle filter and generalized probabilistic data association method was realized in the new algorithm. The rapid expansion of computational complexity,caused by the simple combination of the interacting multiple model and particle filter,was solved by introducing model information into the sampling process of particle state. And the effective validation and utilization of echo was accomplished by generalized probabilistic data association algorithm. The concrete steps of algorithm were given,and the theory analysis and simulation results showed its validity.