系统辨识与建模的一种新方法

A NOVEL ALGORITHM FOR SYSTEM IDENTIFICATION AND MODELING

本文从函数逼近观点研究线性和非线性系统辨识问题 ,导出辨识方程 ,提出用神经网络建立线性和非线性系统的模型 .根据函数内差逼近原理建立神经网络学习方程 ,给出优化算法 .计算机仿真表明新算法计算速度快 ,具有良好的推广。

This paper was a research on the identification of linear and nonlinear systems on basis of function approximation. The identification equation was given. The neural network model of systems have been proposed. The optimal learning algorithm for neural network was constructed. Computer simulation shows that the new algorithm has better approximation, generalization and convergence property.