Autonomous Vehicles with Visual Signals for Pedestrians: Experiments and Design Recommendations

Abstract

Autonomous Vehicles (AV) are the future of transportation and they will transform the dynamic of vehicle and pedestrian interaction. However, in the absence of a driver, it is not clear how an AV can use visual signals to communicate its intention to pedestrians. To explore this question, we conducted four human-subject experiments and evaluated six representative AV visual signals in terms of visibility, intuitiveness, persuasiveness, and usability, thus deriving twelve practical design recommendations for AV visual signals. Moreover, the methodology used in these experiments are novel including one that utilizes a fully automated vehicle. Hence, they enable us to discover new insights about pedestrian crosswalk behaviours and AV-pedestrian trust. So altogether, we hope this work will contribute to the development of various SAE and ISO AV lighting standards, thus taking another step toward the maturation of automated driving.