Classifying Pedestrian Crossing Flows: A Data-Driven Approach Using Fundamental Diagrams and Machine Learning

We investigate the dynamics of pedestrian crossing flows with varying crossing angles to classify different scenarios and derive implications for crowd management. Probability density functions of four key features—velocity, density, avoidance number, and intrusion number—were analyzed to characterize pedestrian behavior. Velocity–density fundamental diagrams were constructed for each crossing angle and fitted with functional forms from existing literature. Classification attempts using avoidance–intrusion numbers and velocity–density phase spaces revealed significant overlaps, highlighting the limitations of these metrics alone for scenario differentiation. To address this, machine learning models, such as logistic regression and random forest, were employed using all four features. Results showed robust classification performance with velocity and avoidance number emerging as the most influential features. Insights from feature importance metrics and classification accuracy offer practical guidance for managing high-density crowds, optimizing pedestrian flow, and designing safer public spaces.

Funding: This research was supported by the National Science Center, Poland through SONATA [Grant 2022/47/D/HS4/02576].

Supplemental Material: The online appendix is available at https://doi.org/10.1287/trsc.2024.0996.