【作者】 郭继丽；

【导师】 宋维公； 王洪瑞；

【作者基本信息】 燕山大学 ， 控制理论与控制工程， 2003， 硕士

【摘要】 耗散系统理论的实质内容是构造一个非负的能量函数，使得系统能量损耗总是小于能量的供给率。事实上，基于李雅普诺夫函数的稳定理论，也可从耗散性的角度加以分析。引入耗散性控制方法可以充分利用物理系统本身的结构特点，为李雅普诺夫函数的构造提供信息，在实际工程中有着广阔的应用前景。本文基于耗散性理论对不确定机器人的控制问题进行了探究。主要给出了两类控制方法，即无源性鲁棒控制和鲁棒控制。无源性理论主要用于分析具有能观测性或能检测性的非线性系统的稳定性问题。稳定性结果由无源系统的稳定性性质得出。在应用无源性理论对机器人系统进行控制的过程中，本文引入了鲁棒KYP引理，简化了控制器的求解过程；并将输入输出线性化法与无源性理论相结合，获得了鲁棒性强，控制精度高的控制器；同时，将微分几何理论推广至机器人系统的控制中，设计出了一种只依赖于不确定性的界和零动态子系统的稳定性的控制器，这种控制器具有结构简单，灵活适用等特点。控制所考虑的主要是干扰抑制问题。在保证系统稳定的前提下，将干扰对系统输出的影响抑制到所要求的最小程度。非线性控制的难点在于求解非线性偏微分方程。针对控制的这一难点，文中给出了三种解决方案。一种是将Backstepping设计方法和控制理论相结合；另外两种是基于机器人标称系统的无源性假设条件，根据系统不确定性所满足的不同条件给出的。这三种控制器的设计都避免了直接求解HJI不等式。更多还原

【Abstract】 The essential content of dissipative systems theory can be mentioned as there is a positive storage function，which guarantees the energy dissipated inside a dynamic system less than the energy supplied from external source. In fact, Lyapunov stable theorems can also be expressed by dissipative theory. By employing dissipative theory, the inherent characters of controlled physical systems may be virtually used to construct a suitable Lyapunov function. Its applications must be far and wide. Based on dissipative theory, this thesis works over uncertain robot systems control and can be divided into two main parts, i.e. passivity robust control and robust control. Passivity systems theory primarily analyzes stability of nonlinear systems under suitable observability or detectability conditions. The stable result is obtained from the stable character of passivity systems. The thesis introduces robust KYP lemma to control robot systems and singles the calculating process. By employ input/output feedback linearization approach and passivity theory at the same time, a strong robust, high accuracy controller is obtained. Furthermore, the differential geometric theory was applied to robot control and a controller which based on the bounded of uncertainties and a stable zero dynamic subsystem is constructed. control is based on the principle of restraining disturbance. Insure the systems’ stable, and make disturbance have a demanded lowest influence to outputs. The difficulty in controlling nonlinear systems using method is lie in calculating HJI equation. To deal with this problem, the thesis provides three schemes. In the first scheme, Backstepping method and control theory are combined; the other two are introduced based on passivity condition of nominal systems and the different conditions of uncertainties. All the three controllers have avoided calculating HJI equation.更多还原