Crowd simulation can be a useful tool for predicting, analyzing, and planning mass-gathering events. The analysis of simulated crowds aims to extract observations to assess occupant interactions and potential crowd flow issues. This paper presents a continuous-space definition of Centroidal Particle Dynamics (CPD) crowd models, then proceeds to present behaviours observed in the simulated crowds. These include organized micro-grouping (flocking), uncooperative behaviors like passage blocking and collisions due to distracted pedestrians. It also briefly explores how spatial design choices could positively impact pedestrian flow. The observations might be of interest to designers of urban and architectural spaces who are looking to improve pedestrian or occupant experience, particularly in high-density crowd scenarios. The presented CPD method is additionally implemented to run on mobile (Android) devices, allowing on-the-field crowd simulation for event planning.

Agent-based pedestrian dynamics, Crowd analysis, Emergent behaviour, Interactive frame-rate, Occupant simulation
9th Annual Symposium on Simulation for Architecture and Urban Design, SimAUD 2018
Department of Systems and Computer Engineering

Hesham, O. (Omar), Princy, (), Aburime, W. (Walter), Rabeh, Z. (Ziyad), Bhushan, S. (Shashi), & Wainer, G.A. (2018). Observed behaviours in simulated close-range pedestrian dynamics. In Simulation Series (pp. 118–125).