基于模糊神经网络的多移动机器人自学习协调系统

研究多移动机器人的运动规划问题．针对机器人模型 未知或不精确以及环境的动态变化，提出一种自学习模糊控制器（FLC）来进行准确的速度 跟踪．首先通过神经网络的学习训练构造FLC，再由再励学习算法来在线调节FLC的输出，以 校正机器人运动状态，实现安全协调避撞．

A SELF LEARNING MOTION PLANNING SYSTEM\=FOR MULTIPLE MOBILE ROBOTS BASED\=ON FUZZYNEURO NETWORKS

In this paper,we discuss motion planning method for multiple mobile robots in a working environment. In order to move safely in the same workspace,robots must be able to track the velocity accurately deduced by its Planning System .Because of uncertainty of the accurate mathematical model and disturbance of the environment ,it is usually difficult to do so. In the worst case, crash between two robots may occur. A Selfadjusting Motion Planning System ,based on FuzzyNeuro networks, is proposed to adjust the actual velocity .Through offline and online learning algorithm,a fuzzy logic controller is constructed to greatly minimize the tracking error, so the safety of the robots can be attained. Simulation results based on the scheme are presented to demonstrate its efficiency.