【作者】 李卫东；

【导师】 钱积新； 马龙华；

【作者基本信息】 浙江大学 ， 控制科学与工程， 2003， 博士

【摘要】 随着科学技术的迅猛发展，信息处理、计算机通讯、机器人控制、生产过程自动化等技术的完善和广泛应用，人们着眼的系统规模越来越大，内容越来越丰富，现象越来越复杂。在通信、制造、交通、军事、生产过程等领域相继出现了一批反映技术发展方向的人造系统，例如通信和计算机网络、计算机集成制造系统(CIMS)、现代交通系统、军事指挥系统、工业生产的控制与调度系统等，这些系统不便于用微分方程或差分方程描述。由于这些系统不同于用微分方程描述的连续系统，也不同于用差分方程描述的离散系统(数字系统)，以往的理论体系已难以适用于这些新的问题，基于对这类系统行为和性能研究的需要，推动了离散事件动态系统(Discrete Event Dynamical System，DEDS)理论的形成和发展。离散事件动态系统(DEDS)理论是一类新的系统理论，反映了各种人造系统运行的内部规律。伴随研究的深入以及重要工程问题和军事问题的促进，导致研究同时包含离散事件过程和连续变量过程的混合动态系统的需要，人们开始认识到一种由离散事件与连续系统交互作用的一类所谓混杂系统问题的重要性。与单一的连续变量动态系统(CVDS)或离散事件动态系统(DEDS)不同，在混杂系统中既包含了服从物理学定律的连续变量动态系统又包含了遵从优化决策信息逻辑原则的离散事件动态系统，而且两者处在一种强相互作用的制约机制中。 本文主要针对一类混杂系统进行控制理论和控制方法的分析和研究。在综合、分析混杂系统现有的控制理论和控制方法的基础上，深入研究了状态反馈控制、神经网络控制等控制策略在混杂系统控制中的应用。采用李雅普诺夫稳定性理论对系统的稳定性作进一步的分析、比较，实现了混杂系统的快速稳定。具体来说，本文的贡献主要在以下几个部分： 1．混杂系统结构分析。从基本组织结构入手，深入分析了混杂系统的概念、性质及特性。按离散事件和连续变量相互作用的类型，对混杂系统进行了分类。采用混杂自动机作为混杂系统的模型，对混杂系统做了深入分析。最后分析了参11 摘要数混杂系统中一类典型的未知参数情况，并给出了确定未知参数的方法。 2．基于李雅普诺夫函数法的混杂系统稳定性分析。分别利用单李雅普诺夫＿函数法和多李雅普诺夫函数法给出了混杂系统稳定的条件，这些条件可表示为线＿性矩阵不等式（LMI）形式，接着给出了基于线性矩阵不等式（LMI）算法确定稳定边界的方法。－ 3．混杂系统的神经网络控制。改进了传统的反向传播神经网络，提出附加动量法的自适应反向传播神经网络算法，并设计了混杂系统的神经网络控制器，‘＿通过神经网络控制的模型预测与反馈校正构成闭环控制系统，实现了混杂系统的＿跟踪控制。 4．混杂混炖系统的状态反馈控制。由单纯的连续混炖系统分析入手，提出＿了基于状态反馈控制的连续混饨系统的李雅普诺夫指数配置控制方法，从而实现系统的稳定性控制。在此基础上，将该控制策略推广到混杂混饨控制系统中，并进一步推导了混杂混炖系统控制器的设计方法，根据非线性系统李雅普诺夫稳定＿性理论，得出了系统的收敛区域。通过对一类混杂混沦系统的控制研究表明，该＿控制方法是有效的＊更多还原

【Abstract】 With the quick development of science and technology,the extensive apply of information handling,computer communicating,robot controlling,and productive process automating,the systems that people meet are large scale,abundant content and complex phenomenon. A lot of man-made systems have appeared which reflect science and technology trend in the area such as communication,manufacture,traffic,military and productive process. It’s not convenient to describe these systems with differential equations or difference equations. These systems are different from not only the continuous systems that are described with differential equations but also the discrete systems that are described with difference equations. The used knowledge is not applicable to these new problems. The demand to study these systems prompts the development of theory of discrete event dynamical systems (DEDS). DEDS theory reflects dynamical law of all kinds of man-made systems. With going deep study,people start knows a few important problems which discrete event interacting with continuous variable that so called hybrid system. Unlike the continuous variable dynamical system (CVDS) or discrete event dynamical system (DEDS),hybrid system contains not only continuous dynamical system subject to Newton’s law of causation but also discrete event system satisfied the optimal decision logic principle,and both of them accord with the strong interaction condition.The main contents contain the analysis and control of hybrid systems. The controls of hybrid systems mainly deal with state feedback control and neural networks control and analysis of hybrid systems mainly deal with theory of Lyapunov stability on the basis of analyzing and synthesizing current control theory and method. In this dissertation,the main contributions are as follows:1. Hybrid system construction analysis. The concept,property and quality of hybrid system got deep into study from basic construction. Hybrid systems areclassified according the interacting type of discrete event and continuous variable. At last,a classic of hybrid system called parameter hybrid system is analyzed and an algorithm is proposed.2. Stability analysis of hybrid system. Two conditions of hybrid system stability are given and an algorithm that determines stable border is given based on linear matrix inequalities.3. Neural networks control of hybrid systems. Proposed a kind of method that is improved back propagation neural network. A neural network controller is designed. The tracking control of hybrid system is realized by model prediction and feedback correct of neural network control.4. State feedback control of hybrid chaotic system. From analyzing continuous chaotic system,proposed Lyapunov exponent control method of continuous chaotic system based on state feedback control. System stability is realized. The control strategy is extended hybrid chaotic system and a designing controller method of hybrid chaotic system is deduced. From Lyapunov stability theory of nonlinear system,got convergent region of system. This control method is demonstrated effective by the study of a kind of hybrid system.更多还原