基于改进粒子群算法的移动机器人定位构图研究

MOBILE ROBOT LOCALIZATION AND COMPOSITION BASED ON IMPROVED PSO ALGORITHM

为深入探究粒子滤波算法,针对粒子群优化算法易陷入局部最优解问题,利用Levy步长对PSO算法的权重和学习因子进行改进,从而改善了对移动机器人位置的最优估计。基于Levy-PSO算法改进粒子群优化的FastSLAM算法,应用matlab软件平台建立地图,构建仿真环境,阐明具体的仿真流程。改进FastSLAM算法和原算法相比,平均相对误差降低了13.5%,证明了改进FastSLAM算法的有效性。通过ROS平台在室内复杂环境开展了建图实验,在建图效果、建图精度以及算法实时性上都有较好的性能指标。通过仿真实验探讨了路标数量与系统性能的关系,以及机器人运动路径与误差消除效果的关系。

Delving into particle filtering algorithms,particle swarm optimization algorithm was prone to local optimal solutions.The Levy step size was used to improve the weight and learning factor of PSO algorithm,thereby the optimal estimation of mobile robot position was improved.Based on FastSLAM algorithm improved by Levy PSO algorithm,the matlab software platform was used to build the map and simulation environment,clarify the specific simulation process.The average relative error of the improved FastSLAM algorithm was reduced by 13.5%,proving the effectiveness of the improved FastSLAM algorithm.The mapping experiments were conducted in complex indoor environments by ROS platform,and it had good performance index in terms of mapping performance,accuracy,and real-time performance.The relationship between the number of landmarks and system performance,as well as the relationship between robot motion path and error elimination effect,was explored through simulation experiments.