메뉴 건너뛰기
소속 기관 / 학교 인증
인증하면 논문, 학술자료 등을  무료로 열람할 수 있어요.
한국대학교, 누리자동차, 시립도서관 등 나의 기관을 확인해보세요
(국내 대학 90% 이상 구독 중)
고객센터 ENG
주제분류

논문 기본 정보

저자정보
(국민대학교) (국민대학교) (국민대학교) (국민대학교)
저널정보
한국자동차공학회 한국자동차공학회논문집 한국자동차공학회논문집 제32권 제12호
오류 신고하기

피인용 0

검색

    초록·키워드

    Various studies have been conducted to develop an autonomous driving system that guarantees the safety of the passengers of an autonomous vehicle. Accurate path prediction is an essential part of improving the autonomous driving system performance to avoid collision and ensure passengers’ safety by determining the risk of collision with nearby vehicles and pedestrians by recognizing various situations. However, it is possible to continue driving with unnecessary and inefficient behaviors due to the inconvenient ride caused by the malfunction of the judgment unit of the autonomous vehicle due to incorrect path prediction and incorrect judgment on the autonomous driving strategy. This study proposes a technique for predicting the path of a vehicle by integrating physical and learning-based models. A deep learning technique using short and long-term memory, a learning-based model that utilizes time series information, predicts the state quantity of the future vehicle based on the past state quantity of the vehicle and calculates the probability for the vehicle’s physical-based model by performing time series classification on CV, CA, CTRV, and CTRA models. The calculated probability for the physical-based model is used as a weight of the trajectory generated by each model later to predict a single path by considering all four physical models.

    최근 본 자료 전체보기

      UCI(KEPA) : I410-151-25-02-091200822