Pedestrian Decision-Making Uncertainty in Urgent Scenarios Modulates Multi-Level, Neural Hierarchies

Uncertainty in pedestrian decision-making in urgent traffic scenarios significantly influences safe interactions with automated vehicles. However, the underlying mechanisms governing such decision-making behavior remain insufficiently understood. To bridge this gap, we design two experimental paradigms of varying complexity to simulate spatiotemporal pressure and the consequences of decision failures: a high-fidelity virtual reality pedestrian-vehicle interaction experiment and a simplified dynamic stimulus task. Our findings reveal that as stimulus urgency increases, humans adjust their decision-making goals, leading to an initial decrease in decision uncertainty, followed by an increase as urgency intensifies. Neurophysiological signal analysis suggests that insufficient perceptual information and evidence accumulation contribute to heightened decision uncertainty. Our study highlights a clear correlation between human decision-making uncertainty and scenario urgency, particularly within a defined urgency range. Integrating foundational cognitive principles and engineering approaches provides insights to address critical safety issues in pedestrian-vehicle interactions within automated transportation systems.