【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.更多还原

【Abstract】 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.更多还原