基于粒子群优化的遗传粒子滤波算法

Genetic particle filter algorithm based on particle swarm optimization

针对遗传粒子滤波算法中粒子匮乏问题,提出一种新的基于粒子群优化的遗传粒子滤波算法。利用粒子群优化算法,驱动粒子向高似然区域移动,以增加有效粒子的数目,从而抑制粒子退化和匮乏现象,同时将遗传算法中的选择、交叉、变异引入粒子滤波,以改善粒子退化及计算量大的问题。实验表明,该算法有效地改善了粒子匮乏现象,同时提高了状态预估的精度,降低了算法的计算量,提高了算法的鲁棒性。

In order to solve the problem of particles scarcity in genetic particle filter algorithm, a novel algorithm based on parti-cle swarm optimization was proposed in this paper. Particle swarm optimization was used to drive particles to high likelihood area. By doing this, the number of effective particles was increased. Thus, particle degradation and shortage were restrained. Simultane-ously, the selection, crossover and mutation of genetic algorithm were introduced into the particle filter for the sake of reducing time-consuming. Experiments showed that our algorithm could improve the lack of particle, the accuracy of state estimates and ro-bustness effectively. At the same time, time-consuming was reduced.