Automated vehicles in urban traffic: the effect of kinematics and eHMIs on pedestrian crossing behavior

Automated vehicles will soon be introduced onto urban roads. To interact with other road users in an expectation conforming way, understanding the role of implicit and explicit communication is inevitable – especially when encountering vulnerable road users. This paper presents a study conducted in a virtual reality pedestrian simulator that evaluates the effects of deceleration and external Human Machine Interfaces on the crossing behavior of pedestrians. Six different approach strategies with two eHMI states were evaluated to evaluate the individual effects of both implicit and explicit communication of yielding vehicles on the crossing initiation times of pedestrians and their acceptance. Results show that explicitly communicating a yielding intent clearly reduces the time for pedestrians to initiate their crossing. However, the distance at which an AV starts to decelerate has a mixed effect on the crossing – while pedestrians tend to prefer early decelerations, these maneuvers have inconclusive effects on the time lost within the interaction.