考察了单输入单输出的Wiener系统、Wiener-Hammerstein系统和多输入多输出的Wiener系统等块结构 (Block-oriented)非线性系统的辨识中辅助变量方法的运用.首先概述已有的块结构系统的辨识方法及其发展现状,以及需要进一步研究的问题.随后分别介绍基于这几类系统的辅助变量辨识方法的基本思想和分析方法,
并运用截尾扩张的随机逼近算法给出具体的递推算法和相关收敛性结果.最后给出仿真例子以说明收敛结果的有效性.

This paper deals with the instrumental variables identification methods for some sorts of block-oriented nonlinear systems: single-input single-output Wiener system, single-input single-output Wiener-Hammerstein system and multiple-input multiple-output Wiener system. First, we put forward the models and gives a general introduction of the development states of these three nonlinear systems identification. Then a specific identification analysis is presented using instrumental variables and the recursive algorithms based on the stochastic approximation with expanding trun- cations are given and its consistent analysis is followed. Two simulation examples are given to verify the theoretical analysis.