智能仓储机器人全局定位方法研究

Global Positioning Method of Intelligent Warehousing Robot

全局定位技术是智能仓储机器人研究的关键环节。针对粒子滤波定位方法存在粒子缺乏导致在未知初始状态时不能进行鲁棒状态精确预估等问题,提出了利用混沌粒子群优化算法对粒子滤波方法进行改进。通过优化,粒子集向后验概率密度分布具有较大取值的区域活动,增加了用于估计的有效粒子,抑制了粒子退化与枯竭,提高了粒子多样性和收敛速度。实验结果表明利用混沌粒子群优化算法改进粒子滤波的方法是有效的,具有较好的鲁棒性和预估精度。

The existence of a lack of particles results in an unknown initial state can not be accurately forecast robust state issues such as targeting methods for particle filter.This paper proposes chaotic particle swarm optimization algorithm particle filter improvements.By optimizing the particle set back the posterior probability density distribution having a larger value of regional activities,it increase the effective particle used to estimate,inhibiting the degradation and depletion of the particle,and improves the particle diversity and the convergence speed.