Research on Model-driven Adaptive Control Systems based on the Error of Closed-loop Parameters

本研究的目的是从控制理论的角度出发，提出一种新的自适应控制方法——“模型驱动控制”，并为控制系统的自适应控制建立理论基础。分析与综合作为强大的控制设计框架。从理论和实验两个方面揭示了模型驱动控制体系结构的意义及其特性。模型驱动控制的一个基本概念来源于人的运动控制。在本研究中，为了在自适应控制和电机控制领域提供处理“复杂系统”方法的新发展，我们主要尝试了以下三个主题：(1)模型驱动自适应控制的建立及其与电机控制的关系；(2)复杂系统辨识方法的建立及其与学习控制的关系；(3)各种非线性复杂系统的控制理论方法。我们提出了一个著名的“反馈误差学习方法”作为二自由度控制系统，并研究了其闭环结构。从理论上证明了闭环系统具有自适应的稳定性和自适应参数的收敛性。进一步，将这些结果推广到反馈回路中存在时滞的情况。(2)中的每个结果都包含了处理复杂系统的基本困难的兴趣方法。在(3)中，我们获得了与各种复杂系统相关的各种成果，例如非线性机械臂、燃料电池中的听觉交换器、控制生物学、人体运动控制的生物力学、量子控制系统、大规模网络系统等。特别是控制生物学和量子控制系统的研究成果，将控制理论的新领域推向了前沿。

The purpose of the research is to formulate a new method of adaptive control called "Model-driven control" from a control theoretical point of view and to establish a theory for. analysis and synthesis as powerful control design framework. We have revealed significances and its properties of model-driven control architecture from theoretical and experimental aspects. A fundamental concept of model-driven control comes from a motor control of human. In this research, in order to provide new developments of method for dealing with "Complicated System" for both fields of adaptive control and motor control we tried the following three themes mainly (1) The formulation of model-driven adaptive control and the relation to motor control, (2) The establishment of identification method for complicated systems and the relation to learning control and (3) Control theoretical approaches to various nonlinear complicated systemsIn (1), we formulated a well-known "Feedback Error Learning method" as two-degrees of freedom control systems and investigated its closed-loop structure. A stability of the closed-loop system including adaptation and a convergence of its adaptation parameters were shown theoretically. Furthermore, these results were extended to a case where a time delay exists in the feedback loop. Each, result in (2) includes interest approaches to the essential difficulties for dealing with complicated systems. In (3), we obtained various outcomes which related to various complicated system, e.g., nonlinear robot arm, a hear exchanger in a fuel cell, control biology, biomechanics of human motor control, quantum control systems, a large scale network system and so on. In particular, results on control biology and quantum control system moved forward the frontier of new field in control theory.