摘要
将强化学习引入机械臂的避碰问题研究,建立了平面三自由度机械臂的多Agent避碰系统,系统结合了最近障碍物信息和偏差角信息来产生控制指令。采用基于K-均值聚类的强化学习方法作为基本的控制策略,给出了系统算法的具体实施过程。通过仿真试验,证明了基于聚类划分的强化学习方法在机械臂避碰问题中的可行性和有效性。
This paper reports on the obstacle avoidance problem for robotic manipulators.The reinforcement learning（RL） method was applied to obstacle avoidance problem and a multi-agent system was built.According the real-time demand of manipulator control, the Sarsa（A） algorithm, which was combined with K-means clustering algorithm, has been selected for its on-policy feature and efficiency. The implement process of the algorithm was given and in the end of this paper, a simulation experiment with different environment was done, the result showed the RL method＇s feasibility and availability.
出处
《机械设计与制造》
北大核心
2007年第8期140-142,共3页
Machinery Design & Manufacture
基金
国家民用航天科研专项计划(科工技[20041530)
关键词
强化学习
避碰
AGENT
K-均值聚类
Reinforcement learning
Obstacle avoidance
Agent
K-means clustering
