Research on Driver Identification Method Applying Conventional Neural Network on Driving Behavior at Intersections

Abstract

Recently in Japan, due to the aging population, there tends to be an increase in traffic accidents involving the elderly driver. The decline in physical, cognitive, and judgment abilities due to aging contributes to the increase in accidents involving elderly drivers. The technology development of autonomous driving has the potential to compensate for these issues. However, application to general roads is still considered to take a lot of time. Thus, it is essential to develop technologies to help in preventing traffic accidents. This technology requires the vehicle to identify (recognize) the driver. In this study, we construct a method for identifying drivers based on their driving performance. We focused on drivers driving behavior at the intersection because it is complicated to operate through intersections and driving characteristics are remarkably visible. In conclusion, in this study, the proposed method applying Conventional Neural Network (CNN) could identify drivers about high accuracy by drivers.