Using Mixed Reality Stated Choice Experiments to Measure the Interactions Between Pedestrians and Autonomous Vehicles

The integration of autonomous vehicles into urban traffic systems raises critical questions about pedestrian decision‐making in mixed environments. Traditional virtual reality methods, while safe and controllable, often lack ecological validity due to their fully virtual settings and absence of real‐world consequences. This study presents a Mixed Reality (MR) methodology, highlighting the engineering design and implementation of a comprehensive experimental system that overlays virtual autonomous vehicles onto real street environments. The principal innovation lies in the end‐to‐end experimental design and physiological measurements, which balances environmental authenticity, participant safety, and precise manipulation of vehicle behaviors and human–machine interface. Participants engaged in road crossing tasks across varied scenarios of vehicle approaching dynamics, which generates movement decisions, emergency indicators, and responses to eHMI messages. Self‐report assessment of engagement, comfort, and immersion experience were also collected. Results confirmed that MR experimental design effectively elicits realistic pedestrian responses while maintaining safety and comfort.