一种用于机械手控制的在线快速学习方法

A Fast Online Learning Method for Manipulator Control

本文在对ＣＭＡＣ网络进行研究比较的基础上，提出了一个用于预定运动轨迹的二关节平面机械手控制的在线直接存贮更新学习方法．模拟实验表明，该学习方法具有较ＣＭＡＣ网络更快的学习收敛速度及更高的控制精度，更少的计算量，且便于硬件实现．

Based on the study of CMAC net, the paper presents a online direct memory refresh learning Method for the online control of the predefined moving route two-joint plane manipulator. The computer simulation result shows that the learning method has faster convergent speed, higher control 'precision and fewer computation, and also more convinient to be realized with hardware.