Formulation of Model Predictive Control Achieving True Eco-Driving of Self-Driving Vehicles

2023 Fiscal Year Final Research Report

• Project/Area Number: 21K14187
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 21040:Control and system engineering-related
• Research Institution: Kyushu University
• Principal Investigator: YUNO Tsuyoshi 九州大学, システム情報科学研究院, 助教 (10756232)
• Project Period (FY): 2021-04-01 – 2024-03-31
• Keywords: 自動車 / 自動運転 / 最適制御 / モデル予測制御 / 省エネルギー

Outline of Final Research Achievements

Development of fuel-economic self-driving vehicles is desired for mitigating traffic accidents, environmental problems, etc. Although the model-predictive eco-driving control have recently attracted great attention, its formulation methodology must be clarified. In this study, we attempted to derive a theoretical foundation of the clarification and to give its extension and practical application. As a result, we obtained important insights into the setting of cost functional and horizon length. Moreover, we developed some bases of their extensions to complicated models and applications to practical problems.

Free Research Field

制御工学

Academic Significance and Societal Importance of the Research Achievements

近年注目を集めているモデル予測制御による省燃費自動運転制御に関して，その評価関数と評価区間の設定に関する重要な知見が得られた．これにより自動運転における燃費性能の向上が見込まれ，自動車の運転コストや環境負荷の低減につながる．また，本研究の問題意識は省燃費運転の本質に迫る極めて重要なものであるにも関わらず，これまで曖昧にされたまま研究が行われてきた．本研究で得られた知見は，省燃費運転制御の分野における基盤を成すものであると考えられる．