Research on Communication Analysis between Vehicles and Pedestrians at Unsignalized Crosswalks Using Online Video

Abstract

In recent years, considerable attention has been paid to research on communication in self-driving vehicles. However, few studies have quantitatively evaluated and clarified the relationship between modern vehicles and pedestrians. Additionally, communication methods and gestures are perceived differently in different countries and regions. In the past, quantitative research on communication that differed across countries and regions often required research teams to travel across countries to conduct research, which entailed enormous costs and a heavy burden. In this study, we propose an online video-based pedestrian behavior analysis and compare pedestrian communication in Japan and U.S. based on a questionnaire survey. Results with the proposed algorithm reveal that Japanese pedestrians tend to use eye contact as a form of communication at unsignalized crosswalks, whereas American pedestrians tend to use hand-raising movements. Moreover, the gesture choices of pedestrians in both countries vary based on the strength of their authority on the road.