Simulating Pedestrian Risk in Dense Urban Networks: Bottlenecks, Flow Interactions, and Intersections

Key Points

• The study aims to quantitatively analyze pedestrian crowd risks in dense urban environments.
• Applied Python-based pedestrian simulation and social force model for risk analysis.
• Executed 24 scenarios combining flow levels, obstacles, and stop probabilities in three urban space types.
• Evaluated risk levels using a composite risk index and validated results with independent random seeds.
• T-junction T08 exhibited the highest risk index (RI = 0.834 ± 0.122).
• X-junction showed the greatest variability (CV ≈ 16.2%), indicating prediction uncertainty.

Abstract