摘要
提出了一种新型的维纳系统,能以任意精度逼近动态非线性对象,证明了此系统在一定条件下,逼近动态非线性系统离散点列的精度优于传统的以神经网络为静态非线性环节的维纳系统.此系统还具有运算量小的特点,易于实现对动态非线性系统进行在线建模与控制。
A novel Wiener-type system is presented and is proved having a fine performance of identifying dynamic nonlinear systems. It can approximate discrete outputs of any dynamic nonlinear systems. We prove, under certain conditions, the novel system has better modeling performance than conventional one whose nonlinear component is also neural network. Meanwhile, the novel system has less computing complexity, so it can model and control dynamic nonlinear systems on line.
出处
《电机与控制学报》
EI
CSCD
北大核心
2004年第2期95-99,共5页
Electric Machines and Control
基金
国家自然科学基金资助项目(60174021
60374037)
南开大学创新研究基金资助项目
关键词
维纳系统
混沌系统
建模
神经网络
非线性动态系统
neural networks
Wiener-type system
chaos identification
dynamic nonlinear system