一种基于动态人工神经网络的Wiener模型辨识  

Identification of Wiener Model Using Dynamic Artificial Neural Networks

作　　者：李世华[1] 吴福保[2] 李奇[2] 

机构地区：[1]东南大学自动控制系,南京210096 [2]东南大学自动化研究所,南京210096

出　　处：《控制理论与应用》2000年第1期92-95,共4页Control Theory & Applications

摘　　要：提出了一种新的辨识模型对Wiener模型进行辨识.该模型由一线性动态神经元串联一静态BP网络模型组成.利用线性动态神经元对Wiener模型的线性动态部分建模,利用静态BP网络逼近模型的静态非线性部分.并且给出了统一的BP辨识算法.仿真结果表明了该方法的有效性.The problem of identification of a Wiener model is studied.The proposed identification method uses a dynamic neural network (DANN) which consists of a linear dynamic neuron (LDN) in cascade with a static BP neural network (SBP).A unified back propagation algorithm is proposed to estimate the weights and the biases of the LDN and the SBP simultaneously.Numerical examples are provided to show the efficiency of the proposed method.

关 键 词：人工神经网络 系统辨识 WIENER模型 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]