eHMI Safety Alerts for Automated Shuttle and e-Scooter Interaction: A Full-Scale Empirical Study in a High-Fidelity Laboratory

Automated Driving Systems (ADS) have the potential to redefine transport systems, but the safety of their interaction with Vulnerable Road Users (VRUs) remains a critical challenge. Automated shuttles are particularly susceptible to close-proximity encounters with micromobility users, such as escooters, which often travel faster than the shuttle in shared urban environments. This study investigates the perception of safety of such interactions through human-in-the-loop experiments in a high-fidelity indoor testing facility. By carefully managing the safety parameters in a controlled environment, we examined a high-risk overtaking scenario where an e-scooter rider passes and overtakes a slowly moving SAE Level 4 ADSequipped shuttle. In an uncontrolled environment of actual traffic, close cut-in manoeuvres could potentially trigger harsh (or emergency) braking that can cause discomfort or injury to passengers. To communicate the vehicle’s hazard zone to the escooter, we evaluated three External Human-Machine Interface (eHMI) modalities: visual (laser projection), auditory (spoken alert), combined (audiovisual), and a control (no communication). The environmental conditions were varied across lighting (from simulated daylight to night) and noise levels (from simulated traffic noise to silence) to assess the robustness of each modality. Utilising a comprehensive dataset comprising overhead video analytics, eye tracking, vehicle telemetry, and questionnaires, we assessed the interaction safety and perceptions of both the escooter rider and shuttle passengers. The results indicate that while eHMI alerts did not significantly alter objective interaction distances, they notably improved the subjective perception of safety for e-scooter riders and passengers. Multimodal alerts, in particular, demonstrated the highest potential. These findings highlight the importance of human-centric communication systems in Automated Public Transport (APT) to foster safer coexistence with external road users and passengers.