关键词: 移动机器人, 同时定位与地图创建, 粒子滤波器, 遗传算法

Abstract: In order to solve the problem that simultaneous localization and mapping (SLAM) algorithm based on Rao-Blackwellized particle filters needs a large number of particles and the frequent resampling might lead to the particle impoverishment, this paper proposes a kind of particle filter based on genetic optimization method. In order to combine the particle weight and the diversity of samples, the paper designs an improved genetic algorithm to replace resampling. It takes into account both the odometer and the observed information when computing the proposal distribution, through the genetic algorithm to keep the diversity of samples. The experimental results show that the new particle filter algorithm performs well on both estimation accuracy and consistency as well as builds maps with higher accuracy.

Key words: mobile robots, simultaneous localization and mapping, particle filter, genetic algorithm