摘要
本文在将文 [6]的参数学习算法推广到多变量系统和对爬山法加以改进的基础上 ,提出了一种新的基于Pi sigma混合型自适应模糊神经网络的多变量自适应控制器 .该控制器能在不需过多先验知识的情况下在线自学习前件和后件参数 .仿真结果表明 。
In this paper, the learning algorithm in paper [6]is extended fo multivariable system and the hill climbing search algorithm is improved. Furthermore, a novel multivariable adaptive controller based on hybrid Pi sigma neural network is proposed, which can learn the parameters of the IF and THEN part of the rules on line with little prior knowledge.The adaptive controller performs encouraging results in the simulation.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
1999年第2期309-312,共4页
Control Theory & Applications
关键词
自适应控制器
自学习
模糊神经网络
控制器
adaptive cntrol
multivariable system
hybrid pi sigma neural network
self-learning