Research of Stop Assistance Considering Visibility of Intersection

Abstract

Crossing collisions caused by cognitive error often occur at unsignalized intersections. For the countermeasure, driver assistant systems that warn a driver against stop sign violation are researched and commercialized. However, if a driver is warned in spite of recognizing an intersection, he or she may feel annoyance and turn the system off. In this research, an index of intersection perception utilizing the saliency map was formulated in order to simulate the driver’s cognition. Then, an assistance method proposed gives a driver early alert only when an approaching intersection is difficult to notice. Applicability of this index was confirmed by a driving simulator, and effectiveness of the proposed assistance method was verified for subjects with the special eyeglasses emulating elderly person’s vision.