【作者】 周颖；

【导师】 武玉强；

【作者基本信息】 东南大学 ， 控制理论与控制工程， 2006， 博士

【摘要】 不确定非线性系统的自适应控制理论及其应用越来越受到人们的重视,现在已经是自动控制研究领域中非常活跃的部分。在现代控制理论的研究中,几何方法和Lyapunov稳定性理论是研究此类问题最重要的研究工具。许多研究考虑具有参数和结构不确定性的非线性控制系统的设计和分析。对具有参数不确定性的非线性系统,基于Backstepping方法的自适应控制研究已成为现代控制理论的主要热点之一。然而,目前已有成果大部分针对的是线性系统或者是单输入单输出非线性系统,而对多输入多输出不确定非线性系统的自适应控制问题研究较少,基于输出反馈的多输入多输出不确定非线性系统自适应控制问题研究则更少。本文研究了一类具有参数化的严格输出反馈形式的多输入多输出非线性系统的自适应控制问题。针对具有不确定参数、受外源信号干扰等情况下的此类受控系统,分别利用Nussbaum增益、高频增益矩阵分解等方法放宽了以往多输入多输出系统中关于高频增益矩阵的假设条件,基于输出反馈,利用Backstepping方法设计了自适应跟踪控制器。并在输出跟踪问题的基础上,进一步研究了此类受控系统的输出机动问题。本文的主要研究成果如下:一、减弱了以往针对多输入多输出系统要求高频增益矩阵K p(或K pS )正定的假设条件,只需受控系统线性部分高频增益矩阵Bm (或Bm S )为Hurwitz。通过设计状态观测器,实现了对受控系统状态的虚拟估计。利用Backstepping方法,对受控系统设计出了基于输出反馈的自适应控制器,使得跟踪误差能够任意小,且闭环系统所有信号有界。二、当受控系统线性部分高频增益矩阵Bm的顺序主子式符号已知时,对高频增益矩阵进行所谓的S1 , D1 ,U 1分解,并设计状态观测器,对受控系统状态进行虚拟估计,利用Backstepping方法,给出了基于输出反馈的自适应控制器设计,使得跟踪误差趋近于零,且闭环系统所有信号有界。三、当受控系统线性部分高频增益矩阵Bm能够转化成为上三角形矩阵或下三角形更多还原

【Abstract】 The problem of adaptive control and application for uncertain nonlinear systems has attracted increasing attention. In the research of modern control theory, geometric method and the Lyapunov-based stability theory are the most important tools to investigate the problem of the adaptive control for uncertainty nonlinear systems. In the design and synthesis of nonlinear systems with parametric and structural uncertainty, especially with parameter uncertainty, the adaptive control based on Backstepping approach has made a breakthrough and become the highlight in the control theory. However, what the most existing results deal with is the SISO nonlinear systems, the results for MIMO nonlinear uncertain systems are seldom found.The goal of this paper is to investigate the adaptive control based on output feedback for a class of parametric strict-feedback multi-input multi-output nonlinear systems with parametric uncertainty. For such MIMO nonlinear systems with parametric uncertainty or exogenous signals, the assumptions for the high frequency gain matrix of the systems are relaxed using so-called Nussbaum gain or the factorization of the high frequency gain matrix. Enlightened by the output tracking design, the output maneuvering for such systems is taken into account. The main results in this paper are as follows:1. Compared to the positive define assumption for the high frequency gain matrix K p (or K pS ) of the MIMO linear systems, in this paper, only the hurwitz is required. The output feedback adaptive controller is presented based on Backstepping method which renders the tracking error arbitrary small and all signals of the closed-loop systems are bounded through the proposed scheme.2. When the sign of the leading minitor of the high frequency gain matrix Bm is known, the output feedback adaptive controller based on Backstepping method is presented using the so-called S1 , D1 ,U 1 factorization for Bm . The design scheme guarantees that the tracking更多还原