• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to previous page

Optimal control technology for complex systems by applying machine learning

Research Project

Project/Area Number 18K04222
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Review Section Basic Section 21040:Control and system engineering-related
Research InstitutionToyota Central R&D Lab., Inc.

Principal Investigator

Ito Yuji  株式会社豊田中央研究所, 量子デバイス研究領域, なし (10613565)

Co-Investigator(Kenkyū-buntansha) 藤本 健治  京都大学, 工学研究科, 教授 (10293903)
Project Period (FY) 2018-04-01 – 2022-03-31
Project Status Completed (Fiscal Year 2021)
Budget Amount *help
¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2020: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Fiscal Year 2019: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Fiscal Year 2018: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Keywordsデータ駆動型制御 / 確率最適制御 / 機械学習
Outline of Final Research Achievements

This study aimed to develop fundamental control theories for data-driven models that can represent complex systems. For data-driven models, such as Gaussian processes, and nonlinear stochastic systems that can express nonlinearity and uncertainty, we developed methods to analyze stability and/or design controllers improving statistical control performance. In addition, we developed methods for stability and/or control of linear stochastic systems that can express various uncertainties, which are expected to be extended to nonlinear stochastic systems.

Academic Significance and Societal Importance of the Research Achievements

機械学習の発展によりガウス過程等のデータ駆動型モデルが注目されており、そのようなモデルを用いて制御則を設計する試みがされている。しかしながら、データ駆動型モデルは高い表現力と代償に複雑な関数で構成されているため、従来の制御理論を用いても統計的最適性や安定性を保証した制御則を設計する事は困難である。本研究で構築した理論はこの問題を解決するための手法であると共に、今後の更なる発展が期待される。また、個体差が大きく外乱にも弱いナノスケールデバイス等、非線形特性や不確かさ(ばらつき)をあわせ持つ複雑な対象を制御する要求に対して、本研究成果の活用が期待される。

Report

(3 results)
  • 2021 Final Research Report ( PDF )
  • 2019 Research-status Report
  • 2018 Research-status Report
  • Research Products

    (5 results)

All 2020 2019

All Journal Article (1 results) (of which Peer Reviewed: 1 results,  Open Access: 1 results) Presentation (4 results) (of which Int'l Joint Research: 3 results,  Invited: 1 results)

  • [Journal Article] Stochastic Optimal Control to Minimize the Impact of Manufacturing Variations on Nanomechanical Systems2019

    • Author(s)
      Ito Yuji、Funayama Keita、Hirotani Jun、Ohno Yutaka、Tadokoro Yukihiro
    • Journal Title

      IEEE Access

      Volume: 7 Pages: 171195-171205

    • DOI

      10.1109/access.2019.2955697

    • Related Report
      2019 Research-status Report
    • Peer Reviewed / Open Access
  • [Presentation] Stability Analysis for Linear Systems with Time-Varying and Time-Invariant Stochastic Parameters2020

    • Author(s)
      Yuji Ito, Kenji Fujimoto
    • Organizer
      21st IFAC World Congress (To appear)
    • Related Report
      2019 Research-status Report
    • Int'l Joint Research
  • [Presentation] Sampling-Based Stability Evaluation with Second-Order Margins for Unknown Systems with Gaussian Processes2019

    • Author(s)
      Yuji Ito, Kenji Fujimoto, Yukihiro Tadokoro
    • Organizer
      2019 IEEE 58th Conference on Decision and Control
    • Related Report
      2019 Research-status Report
    • Int'l Joint Research
  • [Presentation] ガウス過程モデルを用いたフィードバック制御則の設計2019

    • Author(s)
      伊藤優司
    • Organizer
      SICE 制御部門 データ科学とリンクした次世代の適応学習制御調査研究会 第3回講義会
    • Related Report
      2019 Research-status Report
    • Invited
  • [Presentation] On Optimal Control Based on Parametric Gradient Approximations for Nonlinear Systems with Stochastic Parameters2019

    • Author(s)
      Yuji Ito, Kenji Fujimoto, and Yukihiro Tadokoro
    • Organizer
      2019 American Control Conference
    • Related Report
      2018 Research-status Report
    • Int'l Joint Research

URL: 

Published: 2018-04-23   Modified: 2023-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi