On error covariance analysis of an output-error type closed-loop subspace model identification method for unstable systems

2022 Fiscal Year Final Research Report

• Project/Area Number: 18K04217
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 21040:Control and system engineering-related
• Research Institution: Osaka Institute of Technology
• Principal Investigator: Oku Hiroshi 大阪工業大学, 工学部, 教授 (20351455)
• Project Period (FY): 2018-04-01 – 2023-03-31
• Keywords: システム同定 / 閉ループ系 / 不安定系 / 誤差分散解析

Outline of Final Research Achievements

In this study, we analyzed the characteristics of the MOESP-type closed-loop subspace model identification method (CL-MOESP method), which is one of the subspace identification methods based on the output error model in closed-loop. The main research results can be summarized in the following two points. (1) This study gave the theoretical basis that "the CL-MOESP method can be applied to unstable objects to be identified" as opposed to the conventional premise that "objects to be identified are stable". (2) We theoretically clarified the consistency and error covariance analysis of the estimated values of the CL-MOESP method, and quantitatively evaluated the identification accuracy of the method. This indicates that the CL-MOESP method outperforms the prediction-error-based subspace identification methods that have the asymptotical consistency in their estimates.

Free Research Field

制御理論

Academic Significance and Societal Importance of the Research Achievements

今までは相関法を用いた出力誤差モデルに基づく部分空間同定法の漸近特性の解明は未解決であった．しかし，本研究を通してCL-MOESP法の推定値の一致性が示された．これは予測誤差型部分空間同定法がもつ漸近的一致性よりも強い性質であり，CL-MOESP法の持つ優位性の一つである．また，CL-MOESP法の誤差分散解析に関する一定の成果が得られ，将来的にCL-MOESP法から得られる同定モデルの定量的な精度評価法の開発につながり得ると考えられる．最後に，CL-MOESP法が不安定な同定対象にも適用可能であることの理論的保証が得られたため，今後本手法が広く実システムへ適用されることが期待できる．