Crosswalk cooperation: A phone-integrated driver-vehicle cooperation approach to predict the crossing intentions of pedestrians in automated driving

Abstract

While the implementation of automated driving under well-defined circumstances such as highways is possible today, more complex and dynamic environments such as urban areas remain challenging. In particular, the behavior prediction of other road users like pedestrians can be difficult for automated vehicles. We suggest a cooperative approach to overcome this system limitation. More specifically, the system can rely on the users' interpretation of the situation to clarify ambiguity. This study explores how passengers can be used to disambiguate pedestrian crossing behavior. Because passengers likely use their phones while traveling in an automated vehicle, we displayed a live view of the traffic scenes as well as a cooperation request to ask about the intentions of the pedestrian at the crosswalk. A preliminary evaluation of usability shows that this approach provides promising results. Participants also reported trusting the system and showed willingness to help an automated vehicle.