Development of reference shaping theory for automatic driving control systems

2019 Fiscal Year Final Research Report

• Project/Area Number: 16H06094
• Research Category: Grant-in-Aid for Young Scientists (A)
• Allocation Type: Single-year Grants
• Research Field: Control engineering/System engineering
• Research Institution: Osaka University
• Principal Investigator: Minami Yuki 大阪大学, 工学研究科, 准教授 (00548076)
• Project Period (FY): 2016-04-01 – 2020-03-31
• Keywords: 自動走行システム / 予測値整形 / 予測ガバナ / 経路追従制御 / ニューラルネット

Outline of Final Research Achievements

This study focuses on technology to realize an automotive robot that moves intelligently while recognizing the surrounding environment. There are three contributions. First, a design method of a signal shaping mechanism, called prediction governor, was developed. The prediction governor shapes predicted reference signals to reduce the influence of prediction errors on automatic driving. In this study, the prediction governor's usefulness was confirmed through lane-keeping control experiments with a miniature scale experimental apparatus. Then, a data compaction method of deep neural networks, which are used in environmental prediction, was proposed. Finally, a method to control the surrounding lighting environment itself and make it work on automotive robots was presented.

Free Research Field

制御工学

Academic Significance and Societal Importance of the Research Achievements

自動走行システムの開発においては，主として，周辺環境の予測技術と自動車の走行制御技術が別々に考えられていることが多い．これに対して本研究のアプローチは，予測ガバナを導入し，予測技術と制御技術を結びつけるものである．本研究では，設計論の構築と検証を行なっており，この成果は，予測と制御の調和という課題に対する重要な知見を与えているといえる．また，環境と自動車の相互作用に注目し，環境を制御して自動車を操るという今後重要となる新しい問題を検討している．これは，理論面と応用面で今後の発展が期待できる．