基于组合式信号的Hammerstein系统两阶段辨识  

Two-stage identification of Hammerstein system based on combined signals

作　　者：俞洋 张明光 李峰 王正群[2] YU Yang;ZHANG Mingguang;LI Feng;WANG Zhengqun(College of Electrical and Information Engineering,Jiangsu University of Technology,Changzhou 213001,China;School of Information Engineering,Yangzhou University,Yangzhou 225127,China)

机构地区：[1]江苏理工学院电气信息工程学院,江苏常州213001 [2]扬州大学信息工程学院,江苏扬州225127

出　　处：《扬州大学学报（自然科学版）》2022年第5期66-71,共6页Journal of Yangzhou University：Natural Science Edition

基　　金：国家自然科学基金资助项目(62003151);江苏省自然科学基金资助项目(BK20191035);常州市科技计划资助项目(CJ20220065).

摘　　要：针对带有测量噪声和时滞的Hammerstein非线性系统,提出一种基于组合信号的Hammerstein系统两级辨识方法.利用高斯信号和随机信号的组合信号将Hammerstein系统的静态非线性模块和动态线性模块进行分离和识别,简化了识别过程.第一阶段利用单位后向算子的特性将时滞状态空间模型转化为输入输出表达式,利用基于高斯信号输入输出数据的相关分析方法识别动态线性块的参数;第二阶段利用辅助模型技术解决中间变量信息的不可测问题,利用导出的辅助模型递归最小二乘法对静态非线性模块的参数进行辨识.仿真结果表明,该方法能有效地识别Hammerstein系统,具有较好的识别精度和鲁棒性.For the Hammerstein nonlinear system with measurement noise and time delay,a two-stage identification of the Hammerstein system based on combined signals is proposed.The static nonlinear module and dynamic linear module of Hammerstein system are separated and identified by the combined signal of Gaussian signal and random signal,which simplifies the identification process.In the first stage,the time-delay state space model is transformed into the input-output expression by using the characteristics of the unit backward operator,and parameters of the dynamic linear block are identified by using the correlation analysis method based on input-output data of the Gaussian signal.In the second stage,the auxiliary model technique is used to solve the unmeasurable problem of intermediate variable information,and the derived auxiliary model recursive least squares method is used to identify the parameters of the static nonlinear module.Simulation results show that the presented method can effectively identify the Hammerstein system and achieve better identification accuracy and robustness.

关 键 词：HAMMERSTEIN系统 时滞状态空间模型 组合信号源 辅助模型 两阶段辨识 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]