Autonomous driving program considering crossing pedestrians during amber

This study proposes the autonomous driving programs at a signalized intersection. The programs are two types. The first type is amber detection type. This type detects the amber by the camera and makes a judgment of stopping before the intersection or entering the intersection. As a prerequisite traffic rule, when the amber is displayed, the vehicle must not enter the intersection. However, it is excluded when the vehicle can not be stopped safely. More than 60% of the incident cases verified by this type were avoidable. Another type is amber preview type. This type based on vehicle-to-infrastructure cooperative system. This type obtains signal information previously and makes a judgment of stopping before the intersection or entering the intersection. As a programmed rule, if the vehicle can not enter the intersection during amber, or if the vehicle can not pass through the intersection while all the signals are red, even if the signal is green signal, the vehicle stops before the intersection. Approximately 90% of the incident cases verified by this type were avoidable.