基于云进化的遗传粒子滤波算法

Genetic Particle Filter Algorithm Based on Cloud Evolution

针对粒子滤波器存在的粒子贫乏问题,提出了一种基于云模型改进的遗传重采样方法。选择操作采用相隔一定代数进行随机采样的方式,防止选择压力过大导致粒子贫化;利用Y云发生器实现变异操作,根据粒子的观测概率自适应控制搜索范围,在现有粒子的附近搜索精良粒子,在提高粒子有效性的同时增加了粒子的多样性。仿真结果表明：改进后的算法有效地解决了粒子的贫乏问题,提高了滤波性能。

To solve sample impoverishment problem in particle filter application,this paper presents a new genetic resampling algorithm based on cloud model.Random sampling algorithm is brought into selection operation,and particles are selected one time after several iterations to solve sample impoverishment problem caused by too much selection pressure.Y cloud generator is used to realize mutation operation and according to the adaptive control searching area of the observation probability of particles,eminent particles near existing particles can be searched,then particles＇ validity and variety are both improved.The experimental result shows that this algorithm has solved the sample impoverishment problem and improved the filter accuracy.