Skip Navigation
Skip to contents

한국자동차공학회

Login

32권 1호 127-136 2024 [한국자동차공학회 논문집 ]

제목 상대 차량 효용 추정 기반 게임이론을 활용한 자율주행차의 차선 변경 알고리즘 개발
분야 차량동역학 및 제어
언어 Korean
저자 유동우(부산대학교), 김주희(창원대학교), 안창선(부산대학교)
Key Words Autonomous vehicle(자율주행차), Lane change strategy(차선 변경 전략), Game theory(게임 이론), Utility estimation(효용 함수 추정), Interaction(상호작용)
초록 This paper presents an interaction algorithm that can be used in lane change scenarios through a game theory based on estimated opposing vehicle utility. General game theory-based interaction algorithms are valuable in implementing interaction mechanisms that consider the utility of related vehicles. However, they may become inefficient due to uncertainties of the utilities of opposing vehicles. The primary concept behind estimating utility is defining it as a function of achieving driving goals and ensuring safety, which are universal considerations in changing lanes. Weights for each objective in the utility function are determined based on the driver’s driving characteristics. The proposed algorithm estimates opposing vehicle utility by adjusting driving characteristics through driving patterns. Finally, the Stackelberg game, which is grounded in updated utility, offers an optimal lane change strategy after considering the opponent\'s real-time actions. The proposed algorithm was then validated, and could provide a beneficial strategy based on safety, traffic flow, and individual driving preferences.
원문(PDF) 다운로드 Journal Site

사단법인 한국자동차공학회

  • TEL : (02) 564-3971 (사무국 업무시간 : 평일 오전 8시~)
  • FAX : (02) 564-3973
  • E-mail : ksae@ksae.org

Copyright © by The Korean Society of Automotive Engineers. All rights reserved.