摘要
以多移动机器人仓储自动化系统(类KIVA system)为背景,提出一种基于遗传算法与Q学习算法的类KIVA机器人分层路径规划算法。使用遗传算法对移动机器人进行静态路径规划,当机器人按照静态路径行驶时,若检测到即将与其他机器人发生动态碰撞,使用Q学习算法进行机器人之间的局部动态路径规划。对遗传算法和强化学习算法分别提出相应的改进,并通过Python对具体算例进行仿真和有效性验证。
Taking multi-mobile robot warehousing automation system as the background,a hierarchical path planning algorithm for similar-KIVA Robots based on genetic algorithm and Q-learning algorithm is proposed.The static path planning of mobile robot is carried out by using genetic algorithm.When the robot travels along the static path,if the robot detects that it is about to have a dynamic collision with other robots,the Q-learning algorithm is used to plan the local dynamic path between the robots.In addition,the corresponding improvements of genetic algorithm and reinforcement learning algorithm are proposed,and the simulation and validity verification of specific examples are carried out by Python.
作者
张艳伟
吴玉婷
Zhang Yanwei;Wu Yuting(School of Logistics Engineering,Wuhan University of Technology)
出处
《港口装卸》
2019年第3期28-32,共5页
Port Operation
基金
JZW2017115-教育部高等学校物流类专业教学指导委员教研项目
w2016062-武汉理工大学教研项目
关键词
类KIVA系统
路径规划
遗传算法
强化学习算法
similar-KIVA system
path planning
genetic algorithm
reinforcement learning algorithm