摘要
针对粒子滤波算法的退化以及粒子多样性减弱问题,设计了一种新的基于自适应遗传算法的粒子滤波算法.该算法首先用粒子的重要性权重来度量其适应度值,依据粒子的适应度值自适应确定粒子进行遗传操作的概率;然后对选出的粒子实施交叉、变异操作;最后重新评估粒子的适应度并进行状态估计.这种可自适应调节概率的遗传操作能对粒子进行移动,从而提升了粒子的多样性,并使得粒子都能分布在状态的后验概率密度分布的周围.实验结果表明,该算法可有效提高非线性系统状态的估计精度,尤其在系统状态发生突变时,可以得到较好的估计精度.
A new particle filter based on the adaptive genetic algorithm was proposed for moving the degeneracy phenomenon and alleviating the sample impoverishment problem in the particle filter. At first, the algorithm used the importance weight of particles to weigh their fitness value and determined the probability of particles to experience genetic manipulation adaptively according to their fitness value. Then, it implemented the crossover and mutation operation to the samples selected. Finally, it weighed the particles again and estimated the state. By using this genetic manipulation which could adaptively adjust its probability, the particles were moved and the diversity was enriched so as to guarantee that particles are distributed around the posterior density distribution of the state.The simulation results show that the proposed algorithm can effectively improve the state estimation accuracy, especially when the state changes abruptly.
基金
国家自然科学基金(61075032)
国家自然科学基金(60705015)
安徽省自然科学基金(090412059)资助
关键词
粒子滤波
遗传算法
粒子退化
自适应
粒子多样性
particle filter
genetic algorithm
particle degeneracy
adaptive
diversity of particle
