부산대학교 도서관

When drivers miss their intended exits, they should keep going and loop back around. However, some drivers perform abrupt lane changes or slam on the brakes on the shoulder of the road, hoping they can still make the turn. Such driving behavior is risky and is the cause of many fatal crashes. Knowledge of confusion zones where most drivers get confused can be inferred and shared with guidance to mitigate collision risk in these zones. In this paper, we focus on this use case. We propose driver support in confusion zones. The proposed method mines the trajectory data of vehicles and identifies turn loops. When turn loops repeat in a region, that region is tagged as the confusion zone. According to identified confusion zones, guidance (e.g., speed and lane change suggestions) is shared with drivers to help them. We tested the feasibility of the proposed approach through a simulation study. Extensive simulations in different settings demonstrate that driver support in confusion zones could decrease the collision risk by approximately 35%.