Exploring Pedestrian Responses to eHMIs in Virtual Reality: Balancing Efficiency and Safety in Interactions with Automated Buses

Automated buses, which have the potential to enhance public mobility, are being deployed globally. However, due to their non-human driving style and lack of traditional driver communication cues (e.g., eye contact and gestures), these vehicles can pose traffic hazards. This study aims to explore pedestrians' information needs regarding external human-machine interfaces (eHMI) when encountering automated buses in ambiguous situations, such as un-signalized intersections. In a virtual reality (VR) experiment, 32 participants navigated a residential area under four eHMI conditions: (1) no eHMI; (2) eHMI indicating the bus is autonomous; (3) eHMI displaying the bus's intent; and (4) eHMI displaying right of way. Participants' crossing behaviors and responses to questionnaires were collected to assess the impact of eHMI on pedestrian-bus interactions. Results showed that eHMI indicating either the bus's intent or priority reduced pedestrians' hesitation and increased their perceived safety. However, eHMI displaying the bus's intent also fostered pedestrian dependence on the interface, potentially leading to risky behaviors. These findings highlight the need for clear, effective eHMI designs and call for further research to fully assess their impact on pedestrian safety.