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

基于闭环参数误差的模型驱动自适应控制系统研究

• 批准号: 14350225
• 负责人: KIMURA Hidenori
• 金额: $ 4.61万
• 依托单位: The University of Tokyo
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (B)
• 财政年份: 2002
• 资助国家: 日本
• 起止时间: 2002 至 2003
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/en/grant/KAKENHI-PROJECT-14350225/
• 关键词: Model-Driven control; Adaptive control; Complicated system; 2-degrees of freedom control system; control biology; quantum control system; モデル駆動型制御; むだ時間

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.

本文的研究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，即“模型驱动控制”，并为自适应控制系统的设计提供理论基础。分析和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自于人类的运动控制。在本研究中，为了给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形式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;各种非线性复杂系统的控制理论方法在（1）中，我们将著名的“反馈误差学习方法”表述为二自由度控制系统，并研究了其闭环结构。从理论上证明了包含自适应的闭环系统的稳定性及其自适应参数的收敛性。此外，这些结果被扩展到的情况下，在反馈回路中存在的时间延迟。（2）中的每一个结果都包含了处理复杂系统的基本困难的有趣方法。在（3）中，我们得到了与各种复杂系统相关的各种结果，例如，非线性机械臂、燃料电池中的热交换器、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推动了控制理论新领域的发展。

项目成果

中山学之, 木村英紀: "脊髄伸張反射機構と小脳の計算モデルを結合した制御系によるロボットの軌道追従制御"電子情報通信学会論文誌DII. J86-D-II,7. 1078-1089 (2003)

Manabu Nakayama、Hideki Kimura：“使用结合脊髓牵张反射机制和小脑计算模型的控制系统对机器人进行轨迹跟踪控制”IEICE Transactions DII，7. 1078-1089 (2003) ）

M.Yanagisawa, H.Kimura: "Recursive estimation methods for discrete systems Part I : Dynamics of Quantum Feedback Systems"IEEE Transact ions on Automatic Control. Vol.48, No.12. 2107-2120 (2003)

M.Yanagisawa、H.Kimura：“离散系统的递归估计方法第一部分：量子反馈系统的动力学”IEEE 自动控制交易。

Y.Oishi, H.Kimura: "Computational Complexity of Randomized Algorithms for Solving Parameter-dependent Linear Matrix Inequality"Automatica. Vol.39. 2149-2156 (2003)

Y.Oishi、H.Kimura：“求解参数相关线性矩阵不等式的随机算法的计算复杂性”Automatica。

H.Kimura: "Integration of Modeling and Control Through Model-driven Control"The 1st SICE Annual Conference. 2677-2678 (2002)

H.Kimura：“通过模型驱动控制实现建模与控制的集成”第一届 SICE 年会。

J.Terashita, H.Kimura: "Robustness of Feedback Error Learning Method with Time Delay"The 1st SICE Annual Conference. 1835-1839 (2002)

J.Terashita、H.Kimura：“具有时滞的反馈误差学习方法的鲁棒性”第一届 SICE 年会。