딥러닝을 활용한 딜레마 존에서의 차량 속도 제어

Abstract

The dilemma zone refers to the area where a vehicle approaching an intersection as the light turns yellow must decide whether tostop or not. Often, the vehicle’s speed is too fast for it to stop at the stop line before the light turns red, but the vehicle cannot completelyclear the intersection before the three-second yellow signal ends. This can lead to a major accident as the departing vehicle has not clearedthe intersection before the opposite light changes to a green signal and cars on the other side of the intersection start to move. To preparefor this situation, traffic light patterns were modified using pre-trained data sets, then the traffic patterns at these lights were recorded usingcameras. Autonomous vehicles were introduced into these areas and their speed was controlled based on the recognized traffic lightpatterns. This pa