摘要
提出一种基于动态递归神经网络的自适应控制器 ,该控制器能通过自学习不断进行适应性控制 ,且结构简单 ,易于实现 .其主要特点是能够提供一个跟踪网络来辩识系统模型 ,进而确定控制器的网络参数 ,实现间接自适应神经网络控制 .经过对大量非线性系统的仿真研究 ,证明其具有良好的控制性能 .
An adaptive controller with dynamic recurrent neural network is presented, which can continuously perform adaptive control by means of learning strategy and, meantime, is of simple configuration and easy realization of the control. Its major feature is that it can provide a tracing network to identify system model so as to determine the network parameters of the controller and realize an indirect adaptive neural network control. It is proved through a large amount of nonlinear system simulation that this controller possesses a fine control performance.
出处
《甘肃工业大学学报》
北大核心
2001年第3期64-67,共4页
Journal of Gansu University of Technology
关键词
非线性系统
动态递归神经网络
系统辩识
自适应控制
控制器
nonlinear system
dynamic recurrent neural network
system identification
indirect adaptive control
