Abstract: A learning control system of robots with the parallel control of neural network and PD was presented in this paper. A learning algorithm for analog compound orthogonal neural network was given on the basis of the digital compound orthogonal neural network to obtain a high speed-learning algorithm. The simulation results of the robot manipulator with two joints prove that in the tracking desired path lines of the robot, the control method are superior to the conventional control methods in system responses, tracking accuracy, and robustness. The position tracking control obtains satisfactory effectiveness. The analog neural controller provides a novel approach for uncertain control systems.

摘要： 提出一种神经网络与PD并行控制的机器人学习控制系统。为了加快神经网络的学习算法，本文在数字复合正交神经网络的基础上给出一种模拟复合正交神经网络的学习算法，以两关节机器人为对象仿真结果表明，该控制方法使机器人跟踪期望轨迹，其系统响应、跟踪精度和鲁棒性优于常规的控制方法，位置跟踪获得了满意的控制效果。该模拟神经控制器为不确定系统的控制提供了一种新的途径。