Crowd Health Encoding, for Crowd Simulations Using the Smoothed Particle Hydrodynamics Computational Method

Crowd injury modelling focuses on investigating dangerous scenarios involving crowds, including incapacitating, altering behaviour of, or otherwise disturbing individuals in a way which affects the simulation’s motion and status. In this paper, multiple simulated crowd properties are combined into a single model, and a subsequent prototype is designed and produced using recent methods of crowd simulation, to investigate the usage of persistent, proactive, simulated crowd monitoring of dangerous scenarios, for providing crowd management studies with additional safety information about potentially hazardous pedestrian crowd gatherings. “Crowd health” is an introduced term for the involvement of properties that affect a person’s, or their neighbours’, health within a crowd. Properties considered are panic, which may affect movement, and crush, which may cause injuries. Results of experimentation inform the selection and variation of parameters to produce characteristics of, and observe changes in, crowd health, while maintaining simulation stability and control using the Smoothed Particle Hydrodynamics technique. Findings include defining a way to modify a recent crowd simulation model to inform safety by reporting dynamic crowd health.