基于组合信号的Wiener-Hammerstein系统辨识
Identification of Wiener-Hammerstein system based on combined signals
周士博 1杨浩 1杨岳松 1李峰 1曹晴峰2
作者信息
- 1. 江苏理工学院电气信息工程学院,江苏常州 213001
- 2. 扬州大学电气与能源动力工程学院,江苏扬州 225127
- 折叠
摘要
针对噪声干扰条件下的Wiener-Hammerstein系统,提出一种基于组合信号的两阶段辨识算法用于辨识Wiener-Hammerstein系统各个环节参数.利用自回归(autoregressive,AR)模型和有限脉冲响应(finite impulse response,FIR)模型分别建立Wiener-Hammerstein系统的输入和输出线性环节,利用多项式模型建立非线性环节.在第一阶段,基于高斯信号的输入和输出,采用相关性分析方法辨识Wiener-Hammerstein 系统中输入和输出线性环节的参数,有效解决了中间变量不可测的问题.在第二阶段,基于随机信号的输入和输出数据,利用递推最小二乘法辨识非线性环节参数.仿真结果表明,提出的两阶段方法能够有效辨识Wiener-Hammerstein系统,与其他辨识方法相比,辨识精度有所提高.
Abstract
For the Wiener-Hammerstein system with noise interference,a two-stage identification algorithm based on combined signals is proposed to identify parameters of each step of the Wiener-Hammerstein system.The autoregressive(AR)model and finite impulse response(FIR)model are used to establish the input and output linear components of the Wiener-Hammerstein system,and the polynomial model is utilized to establish the nonlinear components.The Wiener-Hammerstein system consists of two dynamic linear links and a static nonlinear link in series.In the first stage,based on the input-output of Gaussian signals,the correlation analysis method is utilized to identify the parameters of the input and output linear links in the Wiener-Hammerstein system,which effectively solves the problem of unmeasurable intermediate variables.In the second stage,based on the input-output data of random signals,the recursive least square method is used to identify the nonlinear link parameters.Simula-tion results show that compared with other identification methods,the proposed two-stage method can effectively identify the Wiener-Hammerstein system,and improve the identification accuracy.
关键词
Wiener-Hammerstein系统/组合式信号/相关分析法/递推最小二乘法Key words
Wiener-Hammerstein system/combined signals/correlation analysis/recursive least squares method引用本文复制引用
基金项目
国家自然科学基金(62003151)
江苏省自然科学基金(BK20191035)
江苏省高等学校"青蓝工程"项目()
常州市科技计划(CJ20220065)
江苏省研究生实践创新计划(SJCX23_1614)
江苏理工学院研究生实践创新计划(XSJCX23_19)
出版年
2024
NETL
NSTL
万方数据