Hit and Run: Evaluating Interface Placements for Automated Vehicle-Runner Interaction

Abstract

AutomotiveUI research has increasingly explored interfaces for automated vehicles (AVs) to explicitly communicate their yielding intentions and resolve space-sharing conflicts with surrounding road users. However, the focus was primarily on walking pedestrians, overlooking runners, who have distinct requirements. In our outdoor study, participants used an augmented reality simulator to navigate a virtual crossing while running. An AV approached and communicated its intentions using red (not-yielding) / green (yielding) signals across: No-Interface, Smartwatch, eHMI, and Road Projections. No-Interface consistently underperformed. The Smartwatch required sustained attention, distracting runners from the road and hindering their movement. The eHMI required the closest proximity to the vehicle, resulting in delayed interpretation and overreliance on the display vs vehicle behaviour. In contrast, Road Projections were viewable through quick glances without diverting attention; runners still considered AV behaviour, mitigating risky crossing decisions. Our results are crucial for promoting public health and AV interface inclusivity beyond walking pedestrians.