Abstract: This paper gives an improved particle filter based on the GA’s resampling in order to solve the problem of particle’s
regression and scarcity.In this method，the particle’s set which is produced by resampling process experiences the GA’s operator，
for example，selection，crossing，mutation and so on.By the GA’s iteration，more better particles are produced and the particle’s
set variety is realized while the particles which have more important weigh are retained.Compared with the traditional
particle filter，this improved method can evaluate the state and track target more accurately.At the same time，it needs few particles.Experimental results show that the method is correct and has practical value.

摘要： 针对普通粒子滤波存在的粒子退化和匮乏缺陷，提出了一种利用遗传算法进行重采样的粒子滤波改进方法。该方法通过对每个采样时刻生成的粒子集合进行选择、交叉和变异等遗传迭代，在现有粒子个数范围内生成更多优良粒子，在保留高适应度粒子基础上实现了粒子集合的多样性。相对于普通粒子滤波，基于遗传重采样的粒子滤波仅需要较少的粒子就可以实现状态的精确估计和目标跟踪。数学方程和序列图像实验结果表明了算法的正确性和实用性。