一种基于CMAC网络的自学习控制器

A New Self-learning Controller Based on CMAC Network

通过分析 CMAC神经网络的学习机制和连续搅拌反应釜的结构 ,提出了一种自动选择学习率的 CMAC自学习控制方法 .给出了自学习控制器的结构和算法 ,并以连续搅拌反应釜模型为对象进行了仿真研究 .这种网络每次学习少量参数 ,算法简单 .仿真结果表明所提出的控制器优于传统的

A new type of CMAC self leaning controller,which can select its own learning rate,has been presented by analyzing the learning principle of CMAC network and CSTR plant.Algorithm has been presented.Simulation study has been carried out basing on the CSTR model.This network only needs to compute few parameters each time and its algorithm is very simple.Simulation results show that this controller is superior to the traditional PID controller.