摘要
本文利用神经辨识器和神经控制器相结合的思想提出了机器人动力学模型完全未知情况下的一种新的学习控制策略．文中描述了机器人系统新的控制结构，并推导出该结构下神经辨识器和神经控制的在线学习算法．仿真结果表明，文中提出的控制方法能够实现任何可达期望曲线的机器人高精度轨迹控制．
This paper presents a new learning control method on the robot trajectories tracking of unknown model system based on the both the inverse dynamics neural network identifier and feedforward compensation network controller.Basic control configuration are briefly presented and new on-line training shemes are proposed.Simulation results show that it can follow any arbitrarily prescribed trajectories.
基金
国家教委留学回国人员科研启动基金
