基于自适应遗传算法的粒子滤波器

Particle filter based on adaptive genetic algorithm

针对重采样导致的权值退化问题,应用遗传算法的进化思想来优化重采样算法,将粒子权值作为适应度值,合理设定阈值,利用最佳个体保存法保存高适应度粒子,利用自适应交叉、变异操作对低适应度粒子进行进化,将高适应度粒子与进化粒子组合成新的粒子集进行状态估计。仿真实验表明,该算法具有良好的实时性和估计精度,其状态估计精度比标准粒子滤波提高近24倍,比无迹卡尔曼粒子滤波提高近4倍,耗时约为无迹卡尔曼粒子滤波的1/10。

An improved adaptive genetic particle filter algorithm is proposed in order to alleviate weights degradation of particle filtering algorithm.Particle weight is regarded as fitness values,and a percentage of big weight particles are obtained with the best individual preservation method.Crossover and mutation operations are adopted for the remaining particles.Then formed a new set of particles with saved particles,crossover and mutation particles,and state estimation calculations is done.Maintaining the diversity of the particles at the same time,it avoids algorithm falling into local optimum and improves the global search ability of the algorithm.The simulation results show that,compared with the standard particle filter,the proposed algorithm can improve the accuracy of state estimation by nearly 24 times,4 times higher than that of the Kalman particle filter,and it has high real-time performance and good estimation accuracy.