Identification of Hammerstein and Wiener Models Using Spectral Magnitude Matching
Identification of Hammerstein and Wiener Models Using Spectral Magnitude Matching
复制标题
使用谱幅度匹配识别 Hammerstein 和 Wiener 模型
DOI：
10.3182/20080706-5-kr-1001.01086
复制
发表时间：
2008
期刊：
影响因子：
--
通讯作者：
L. Gan
中科院分区：
文献类型：
--
作者：
E. Abd;L. Gan
本文研究了Hammerstein模型和Wiener模型的辨识问题。本文提出的方法利用频谱幅度匹配法，最小化非线性系统的测量输出信号和非线性模型的输出信号的频谱幅度之间的和平方误差。利用广义牛顿迭代算法对Hammerstein和Wiener模型的系数进行了估计。仿真结果表明，该方法在中高信噪比情况下取得了很好的效果。
Abstract The problem of the identification of Hammerstein and Wiener models is considered in this paper. The suggested approach in this paper utilizes the spectral magnitude matching method that minimizes the sum squared error between the spectral magnitudes - evaluated for a number of short-time frames - of the measured output signal of the nonlinear system and the output signal of the nonlinear model. The coefficients of Hammerstein and Wiener models are estimated using the generalized Newton iterative algorithm. Simulation results show that the suggested approach gives very good results especially for moderate and high signal to noise ratios.