Development of System Identification Device for Control Systems Design

控制系统设计系统辨识装置的开发

• 批准号: 12555121
• 负责人: ADACHI Shuichi
• 金额: $ 5.18万
• 依托单位: Utsunomiya University
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (B)
• 财政年份: 2000
• 资助国家: 日本
• 起止时间: 2000 至 2002
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/en/grant/KAKENHI-PROJECT-12555121/
• 关键词: Contorol; Modeling; System Identification

This research develops a system identification device for control system's design. A prototype of the hardware has been developed and a GUI (Graphical User Interface)that supports system identification operation has been installed in the system identification device. The system identification device also includes new system identification methods such as1) System identification method based on SVM (support vector machine),2) System identification methods for non-uniform sampling data.The former method that is SVM-based system identification method is proposed for linear regression models such as ARX (Auto-Regressive with eXtra input) model, and it is shown that the method is robust for input-output data with outliner.In the latter method, non-uniform sampling data is first transformed to uniform sampling data whose sampling period is smaller than the original data. Second a system identification method with missing data estimation is applied to the data.The effectiveness of the developed system identification device is examined using experimental apparatus such as an inverted pendulum and an active noise control svstem.

本研究发展一种系统辨识装置，以供控制系统设计之用。硬件的原型已经开发和GUI（图形用户界面），支持系统识别操作已安装在系统识别设备。系统辨识装置还包括新的系统辨识方法，如1）基于SVM的系统辨识方法（2）非均匀采样数据的系统辨识方法，前一种方法是针对线性回归模型（如ARX）提出的基于支持向量机的系统辨识方法（Auto-Regressive with eXtra input）模型，并证明了该方法对具有轮廓线的输入输出数据是稳健的。首先将非均匀采样数据变换为采样周期小于原始数据的均匀采样数据。其次，将一种具有缺失数据估计的系统辨识方法应用于数据，并利用倒立摆和有源噪声控制系统等实验装置验证了所开发的系统辨识装置的有效性。

项目成果

S.Adachi ほか4名: "System identification and physical parameter estimation of anti-vibration units in semiconductor exposure apparatus"IFAC SYSID'01. (2000)

S.Adachi 等 4 人：“半导体曝光设备中抗振单元的系统识别和物理参数估计”IFAC SYSID01 (2000)。

足立, 高梨, 山口: "不完全な情報下におけるシステム同定法"電気学会論文誌(D). 123-D(発行予定). (2003)

Adachi、Takanashi、Yamaguchi：“不完全信息下的系统识别方法”，日本电气工程师学会会刊（D）（待出版）。

S.Adachi, T.Ogawa: "A new system identification method based on support vector machines"IFAC Workshop on Adaptation and Learning in Control and Signal Processing. (2001)

S.Adachi、T.Okawa：“一种基于支持向量机的新系统识别方法”IFAC 控制与信号处理适应与学习研讨会。

S.Adachi, T.Ogawa, R.Konno: "A system identification method for linear regression models based on support vector machine"Trans.of SICE. 37. 1189-1191 (2001)

S.Adachi、T.Okawa、R.Konno：“一种基于支持向量机的线性回归模型的系统辨识方法”Trans.of SICE。

S. Adachi, T. Ogawa, R. Konno: "A system identification method for Linear regression models based on support vector machine"Trans. of SICE. 37・12. 1189-1191 (2001)

S. Adachi、T. Okawa、R. Konno：“基于支持向量机的线性回归模型的系统识别方法”，SICE 37・1191（2001）。