摘要
针对移动机器人导航控制中的行为协调问题,提出了一种学习速率可控的学习自动机。该方法将机器人与障碍物之间的接触时间变化作为奖惩信号,通过主动控制机器人线速度来调节学习决策时间,根据环境动态特性调整行为动机,并通过学习决策时间和行为动机控制共同控制学习速率,保证机器人在学习决策时间内完成导航行为的协调执行。仿真证实提出的学习自动机应用于移动机器人导航控制是可行的,与动力学分岔控制方法对比,在未知动态环境中进行导航行为协调控制,提出方法的安全性更高。
With an aim at behavior coordination problem in the mobile robot navigation control, this paper suggests a kind of learning automata with the controllable rate.The learning automata uses the contact time variation between the robot and obstacles as the reward and penalty signals to regulate the learning and decision time through the active control over robot linear velocity and to adj ust the behavior intension in terms of environ dynamic behaviors and to control over the shared management learning rate through the learning decision time and the behavior intension control,whereby ensuring that the robot can complete the coordination implementation of naviga-tion behaviors within the learning decision time.The simulation results indicate that the learning automata suggested in this paper in applying to the navigation control by mobile robot is feasible, and in comparison with the dynamic bifurcation control method,its safety in unknown dynamic environ to carry out navigation behavior coordination control is much higher.
出处
《西安理工大学学报》
CAS
北大核心
2015年第3期310-315,共6页
Journal of Xi'an University of Technology
基金
国家自然科学基金资助项目(10872160
51275407
51475365)
陕西省自然科学基础研究计划重点资助项目(2011JZ012)
关键词
移动机器人
行为动力学方法
行为协调
学习自动机
mobile robot
behavior dynamics method
behavior coordination
learning automata
