一种迷宫机器人的人工脑系统  

An Artificial Brain System of Maze-like Robot

作　　者：阮晓钢[1] 许晓明[1] 李欣源[1] 

机构地区：[1]北京工业大学人工智能与机器人研究所,北京100124

出　　处：《控制工程》2010年第5期686-690,共5页Control Engineering of China

基　　金：国家自然科学基金资助项目(60375017);北京市人才强教计划项目;高等学校博士学科点专项科研基金资助项目(20050005002)

摘　　要：提出一种迷宫机器人的人工脑系统,包括迷宫路标感知单元与行为决策单元。其中,感知单元基于ART1神经网络,用于识别迷宫导向路标;决策单元基于行为概率实现矩阵,并以强化学习更新行动策略。机器人所在迷宫的特征为每个路口设有导向路标,路标为含噪声的符号图像。在仿真实验中,令机器人在迷宫中随机行走,通过调节人工脑系统的试验参数,经过一段时间的自主探索学习过程机器人能最终穿越迷宫。仿真实验结果表明,该人工脑系统能够自组织地理解迷宫中导向路标的含义,并引导机器人成功穿越迷宫。同时,该人工脑系统对于基于路标导航的城市巡逻机器人、高危复杂环境下的抢险机器人的研究发展有一定的推动作用。An artificial brain system of a maze-like robot is presented,which comprises a perception unit and a decision-making unit.The perception module is based on ART1 neural network and trained to identify the signposts of the maze.The decision-making units is based on behavior probability matrix,and uses reinforcement learning to update the action strategy.The maze has signposts at each intersection,which are 2-D symbols with noise.In the simulation tests,the robot moves randomly in the maze.By adjusting the parameters of the experiment,the robot will eventually pass through the maze after a learning process during the self-exploration.The simulation result shows that the artificial brain system can be self-organized to make sense of the signposts and successfully guide the robot through the maze.It is also has a positive impact on patrol robot and rescue robots based on signposts navigation.

关 键 词：ART1神经网络 强化学习 自组织 非监督学习 人工脑 

分 类 号：TP27[自动化与计算机技术—检测技术与自动化装置]