Understanding Pedestrian Behavior in Shared Spaces: A Virtual Reality Study of Automated Shuttle Interactions

Abstract

The integration of automated shuttles into shared urban spaces presents challenges due to the absence of traffic rules and the complex dynamics of pedestrian interactions. This paper presents a Virtual Reality (VR) study simulating diverse pedestrian-automated shuttle interactions in a shared space environment, including approaching from various angles and navigating continuous traffic. We focus on key behaviors, such as hesitation, deviation, gazing, and proxemics, that play a critical role in shaping safe and efficient interactions. Additionally, we propose a hybrid model that combines psychological principles with deep learning for trajectory and intention prediction for enhanced interpretability and accuracy. This approach aims to promote the harmonious integration of automated shuttles into urban environments, ensuring safer and more efficient shared spaces.