Eye Gaze-Informed and Context-Aware Pedestrian Trajectory Prediction in Shared Spaces with Automated Shuttles: A Virtual Reality Study
Researchers have developed a new multi-modal prediction model that fuses eye gaze, head orientation, and situational context to predict pedestrian trajectories around automated shuttles. The study, conducted in a virtual reality environment, found that eye gaze provides valuable predictive information, particularly at acute angles where pedestrians actively track the shuttle. Continuous gaze orientation proved more effective than categorical labels, and combining gaze with contextual information significantly reduced prediction errors. AI
IMPACT Enhances safety and efficiency for autonomous vehicle navigation in shared human-robot spaces.