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

Driving behavior modeling based on stability and safety oriented inverse reinforcement learning

Research Project

Project/Area Number 21H03517
Research Category

Grant-in-Aid for Scientific Research (B)

Allocation TypeSingle-year Grants
Section一般
Review Section Basic Section 61050:Intelligent robotics-related
Research InstitutionTokyo Institute of Technology

Principal Investigator

Shimosaka Masamichi  東京工業大学, 情報理工学院, 准教授 (40431796)

Co-Investigator(Kenkyū-buntansha) 小竹 元基  東京工業大学, 工学院, 教授 (10345085)
Project Period (FY) 2021-04-01 – 2024-03-31
Project Status Completed (Fiscal Year 2023)
Budget Amount *help
¥17,030,000 (Direct Cost: ¥13,100,000、Indirect Cost: ¥3,930,000)
Fiscal Year 2023: ¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2022: ¥6,760,000 (Direct Cost: ¥5,200,000、Indirect Cost: ¥1,560,000)
Fiscal Year 2021: ¥5,980,000 (Direct Cost: ¥4,600,000、Indirect Cost: ¥1,380,000)
Keywords逆強化学習 / 運転行動モデリング / 経路生成 / RRT / 運転行動シミュレーション / 確率的パス探索 / 模倣学習 / 逆最適制御 / 高速パス探索 / 最大エントロピー学習 / パス探索 / 重点サンプリング / 運転行動予測
Outline of Research at the Start

交通事故撲滅に向け先進運転支援システムの技術開発が近年盛んに行われている.一方,生活道路中の死亡交通事故件数の減少率は緩慢であり,更なる知的支援技術の基盤として,熟練ドライバの運転特性を模倣するモデリング手法(行動予測・シミュレーション行動生成)に期待が高まっている.本研究では,逆強化学習を基盤に,既存技術に比べ高い安定性・安全性を指向する方法論の確立を目指す.具体的には1)大域的最適性を追求する探索ベースの予測技術(安定性),2)熟練者の教示(正例)に加えて負例を導入したモデリング(安全性)を構築し,構築した枠組の評価を行う.

Outline of Final Research Achievements

In recent years, the development of advanced driver assistance systems has been active. In this study, we developed a method for modeling driving behavior (driving behavior prediction and simulation generation) based on inverse reinforcement learning (imitation learning and inverse optimal control) with a view to the advancement of these systems. In particular, from the viewpoint of application to modeling of automobile driving behavior, we developed techniques oriented toward stability and safety, which have been lacking in conventional techniques. Specifically, we developed a probabilistic path generation method, an efficient reward estimation method using sampling based on path candidates, and an efficient learning method using negative examples (data corresponding to traffic accidents). In addition, to safely collect data corresponding to negative examples, we constructed a driving behavior data collection environment and actually collected driving data.

Academic Significance and Societal Importance of the Research Achievements

本研究で構築した技術の一つである,RRTに基づく経路生成法,及び,重点サンプリング法に基づく報酬場最適化手法は,逆強化学習に基づく運転行動モデリングの適用範囲を拡張するものである,また,負の事案を用いた手法は,熟練ドライバの運転行動の表層的な模倣だけでなく,安全性という価値感を含めた模倣を目指す基盤となっている.構築した技術は交通事故を低減する新たな運転支援システムの開発に貢献する.例えば,事故要因分析や事故回避のシミュレーションへの応用が期待できる.

Report

(4 results)
  • 2023 Annual Research Report   Final Research Report ( PDF )
  • 2022 Annual Research Report
  • 2021 Annual Research Report
  • Research Products

    (8 results)

All 2024 2023 2022 2021

All Journal Article (1 results) Presentation (7 results) (of which Int'l Joint Research: 3 results)

  • [Journal Article] Driving Behavior Modeling in Residensial Roads with Inverse Reinforcement Learning2021

    • Author(s)
      下坂正倫
    • Journal Title

      Journal of the Robotics Society of Japan

      Volume: 39 Issue: 7 Pages: 631-636

    • DOI

      10.7210/jrsj.39.631

    • NAID

      130008083621

    • ISSN
      0289-1824, 1884-7145
    • Related Report
      2021 Annual Research Report
  • [Presentation] Inverse Reinforcement Learning with Failed Demonstrations towards Stable Driving Behavior Modeling2024

    • Author(s)
      Minglu Zhao, Masamichi Shimosaka
    • Organizer
      2024 IEEE Intelligent Vehicle Symposium (IV)
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research
  • [Presentation] 危険行動の時間的局所性に着目した負例導入逆強化学習の安定化2023

    • Author(s)
      趙 茗王路, 羊 少宇, 下坂 正倫
    • Organizer
      ロボティクス・メカトロニクス講演会2023
    • Related Report
      2023 Annual Research Report
  • [Presentation] Improved Template-Based RRT for Efficient Kinodynamic Motion Planning2023

    • Author(s)
      Shaoyu Yang, Minglu Zhao, and Masamichi Shimosaka
    • Organizer
      ロボティクス・メカトロニクス講演会2023
    • Related Report
      2023 Annual Research Report
  • [Presentation] Sequential 2D Continuous Kinodynamic RRT for Driving Behavior at Un-signalized Intersections with Stop Lines2022

    • Author(s)
      Shaoyu Yang and Masamichi Shimosaka.
    • Organizer
      日本ロボット学会
    • Related Report
      2022 Annual Research Report
  • [Presentation] RRT-based maximum entropy inverse reinforcement learning for robust and efficient driving behavior prediction2022

    • Author(s)
      Shinpei Hosoma, Masato Sugasaki, Hiroaki Arie, and Masamichi Shimosaka
    • Organizer
      2022 IEEE Intelligent Vehicles Symposium (IV)
    • Related Report
      2021 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Smooth and Stopping Interval Aware Driving Behavior Prediction at Un-signalized Intersection with Inverse Reinforcement Learning on Sequential MDPs.2021

    • Author(s)
      Shaoyu Yang, Hiroshi Yoshitake, Motoki Shino, and Masamichi Shimosaka.
    • Organizer
      2021 IEEE Intelligent Vehicles Symposium (IV)
    • Related Report
      2021 Annual Research Report
    • Int'l Joint Research
  • [Presentation] 特徴量探索とパラメータ最適化の交互実行によるコンパクトな運転行動推定モデル2021

    • Author(s)
      平川 優伎, 下坂 正倫
    • Organizer
      第22回 計測自動制御学会 システムインテグレーション部門講演会
    • Related Report
      2021 Annual Research Report

URL: 

Published: 2021-04-28   Modified: 2025-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi