未知环境下机器人避障设计研究  

Design of Robot Obstacle Avoidance in Unknown Environment

作　　者：刘欢[1] 王健[1] 李金凤[1] 李金娜[1] 

机构地区：[1]沈阳化工大学,辽宁沈阳110142

出　　处：《机械设计与制造》2013年第10期236-238,共3页Machinery Design & Manufacture

摘　　要：随着信息技术的发展,各个领域越来越需要高性能的自动化系统。机器人技术飞速发展,研究重点已经转向在复杂、未知、不可预测环境中独立工作的自主式智能机器人。介绍了机器人Q学习避障算法的实现方法,并构建了仿真实验平台,模拟了移动机器人在未知环境下自主地、安全地从起始点到达目标点的过程。通过仿真实验验证了Q学习实现机器人在未知环境下的行为选择控制是可行的、有效的,并验证机器人在未知环境下具有良好的越障性能。With the development of information technology, high-performance automation systems are increasingly required in various fields. With rapid development of robotics technology, research focus has been shifted to autonomous intelligent robots working independently in a complex, unknown, unpredictable environment.It describes the Q-learning algorithm to achieve obstacle avoidance methods, and build a simulation platform to simulate the process of a mobile robot reaching the target from the starting point independently and safely in unknown environment. The simulation experiments validate that the control over a robot＇s choice of behavior in unknown environment through Q-learning is feasible and effective, and verify that a robot has a good performance on obstacle crossing in unknown environment.

关 键 词：机器人 Q学习算法 避障 神经网络 

分 类 号：TH16[机械工程—机械制造及自动化] TP242[自动化与计算机技术—检测技术与自动化装置]