基于强化学习算法的井下移动机器人路径规划  

Path Planning of Mobile Robot Based on Reinforcement Learning Algorithm under Coal Mine

作　　者：王帅[1] 

机构地区：[1]煤炭科学研究总院抚顺分院,辽宁抚顺11300

出　　处：《电气技术》2008年第8期47-49,共3页Electrical Engineering

摘　　要：移动机器人路径规划一直是机器人研究领域中的难点问题,本文针对煤矿井下环境的不确定性,采用强化学习算法中的Q-learning算法实现井下移动机器人的局部路径规划,并对Q函数中的即时回报进行加权修正,使算法更有效地利用环境特征信息,提高了避障能力。仿真实验说明该方法的有效性和可行性。Path planning of mobile robot is always a difficult question in the robot research domain. In this paper one of reinforcement learning algorithm Q-learning algorithm is used to realized local path planning of mobile robot under coal mine for environment uncertainty of coal mine. Instant rewards in the function are modified by using weight then the information of environment characteristics is used effectively to avoid the obstacles. Simulation shows that this algorithm is efficient and feasible.

关 键 词：移动机器人 不确定环境 强化学习 路径规划 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]