基于改进遗传粒子滤波的纯方位机动目标跟踪

Bearing-only maneuvering based on improved genetic particle filter algorithm

采用传统遗传粒子滤波算法解决纯方位机动目标跟踪时,由于进化过程中大权值父粒子产生新生代粒子的变异方式比较单一,会导致目标在机动时跟踪误差较大和收敛时间长的问题。为解决该问题,提出一种在进化时调整新生代粒子的分布方式及变异幅度的改进遗传粒子滤波算法,以扩大粒子分布范围、多样化变异幅度并提高粒子接近机动目标真实运动状态的可能性。经仿真实验验证,本文的改进遗传粒子滤波算法能有效提升纯方位机动目标跟踪的收敛速度,减小跟踪误差,提升系统的估计性能。

The distribution of new generation of particles produced by the variation of the dominant parent particles in the bearing-only maneuvering target tracking based on traditional genetic particle filter is single,which will lead to large tracking error and long convergence time.To solve this problem,an improved genetic particle filtering algorithm that adjusts the distribution mode and variation amplitude of the new generation of particles during evolution is proposed to expand its distribution range,multiply the range of variation and increase its likelihood of approaching real state of the maneuvering target.Simulation results show that it can speed up convergence and reduce tracking error,which improves the estimation performance.