【作者】 王仲民； 岳宏；

【Author】 Wang Zhongmin~(1,2), Yue Hong~1(1. Institute of Robotic and Automation, Hebei University of Technology, Tianjin 300130 2. Dept. of Mechanical Engineering, Tianjin University of Technology and Education, Tianjin 300222)

【机构】 河北工业大学机器人及自动化研究所；

【摘要】 将Q强化学习算法应用于移动机器人局部路径规划,解决了移动机器人在复杂环境中的局部路径规划问题。采用基于信任分配的CMAC神经网络实现了该算法,显著提高了传统CMAC在线学习的速度与准确性。仿真实验证明:该强化学习算法不仅能够适应复杂的环境,而且具有较强的自学习能力。更多还原

【Abstract】 In this paper Q reinforcement learning algorithm is adopted for mobile robot local path planning. It makes mobile robot resolve the problem of local path planning in a complex environment. By using CMAC neural network based on credit assignment this algorithm is implemented, and conventional CMAC’s online learning speed and its accuracy are improved at the same time. Simulation experiments prove that this algorithm introduced adapt any complex environment and own good self- learning abilities.更多还原