도심 자율주행차의 횡방향 주행행태 평가 및 시사점 도출 연구

초록

Automated Vehicles (AVs) are designed to interact safely with surrounding objects; however, their responses can be adversely affected by factors such as sensor performance, the responsiveness of driving algorithms to unexpected situations, roadway geometry and facilities, traffic conditions, and weather conditions. To evaluate the lateral driving safety of AVs, this study introduces an assessment metric referred to as the lateral deviation difference (LDD). LDD values in autonomous driving and manual driving modes (AD and MD modes, respectively) were determined when AVs traversed various urban infrastructure sections and compared. Results were classified into three categories: (1) sections where AD mode was significantly safer; (2) sections where AD and MD modes were non-significantly different; and (3) sections where MD mode was significantly better. Cases 2 and 3 corresponded to roundabouts and bus stop zones, respectively, for which countermeasures to enhance AV driving safety are proposed. The proposed LDD metric and evaluation methodology offer a means of identifying road segments that require targeted safety improvements in preparation for the widespread deployment of AVs.

키워드