The auxiliary model identification idea,the multi-innovation identification theory,the hierarchical identification principle, and the coupled identification concept are new methods for studying identification problems, proposed by the author of this article. They have been applied to many identification researches and resulted in different identification method families, and can be used to solve adaptive signal processing, adaptive parameter estimation, a-daptive filtering and prediction, adaptive control and other issues for many linear or nonlinear models. Any objective thing has a dual property: some of the characteristics variables are observable and some are unmeasurable. For the observed system,it is easy to set up the mathematical model for describing its law of motion. The unmeasurable property of thing brings us particular difficulties for setting up the mathematical model. In this case, how to use the measured information of the system to estimate the unknown variables of a system and then to establish the mathematical model of the system is a challenging research topic in the area of system identification, which is the root of the auxiliary model identification idea. This article introduces the auxiliary model identification idea and some auxiliary model based identification methods.%辅助模型辨识思想、多新息辨识理论、递阶辨识原理、耦合辨识概念是该文作者提出的研究辨识问题的原创性新方法,已经被用在很多辨识问题的研究中,形成了不同的辨识方法族,可以用于解决许多线性或非线性模型的自适应信号处理、自适应参数估计、自适应滤波和预测、自适应控制等问题.由于客观事物具有双重属性:一些特征变量是可观测的;一些是不可测的.如果表征系统特征的观测变量都是可测的,就容易建立描述其运动规律的数学模型.客观事物的不可测属性给建立系统数学模型带来特别的困难.在这种情况下,如何利用系统的可测信息,实现对系统未知变量的估算,来建立系统的教学模型,是辨识领域极具挑战性的研究课题.辅助模型辨识思想就是在这样的背景下发展起来的.该文介绍辅助模型辨识思想和一些基于辅助模型的辨识方法.
展开▼
