基于模糊认知图的机器人高层规划研究

本文采用模糊认知图模拟机器人在复杂环境下的动态行为,来对机器人进行高层规划.基于模糊认知图的机器人高层规划系统不仅在一定程度上克服了基于符号机器人规划存在的不适用于含有许多独立子问题或子系统的大系统及机器人之间的合作与协调机制较难寻找等问题,而且还能有效克服基于进化机器人、基于神经网络机器人进行动态行为规划时算法和实现的高度复杂性、需要大量的训练样本及时间、易产生干扰等问题.实验表明把模糊认知图应用到机器人高层规划系统能较好的解决以上问题,且具有简单、鲁棒等特点.

Research on Robot High-Level Planning Based on Fuzzy Cognitive Map

Fuzzy Cognitive Map (FCM) is applied to simulate the dynamic behaviors of the robot in complicated environments and make layout for the robot.It overcomes the defects that are not only symbolic high-level planning unfitness of a big system which has a lot of subsystems and cooperation mechanisms, but also NN and GA in which algorithms are implemented with high complexity and many training examples without cooperation mechanisms. Finally, the experimental results indicate that the robot's high level planning by FCM is more simple, robust and flexible than classical robot's planning.