Summary

Inside Data Science & Artificial Intelligence Labs of THALES, we have been designing a Crowd Simulation engine called SE-Star since 2008.

SE-Star is a life simulator managing: internal variables, motivations & emotions, stimuli, personality and behaviors.

SE-Star uses a Biomimetic control architecture based on navigation (Hippocampus), action selection (Basal ganglia) and planning (Frontal cortex).

Use-cases are: design of critical infrastructures and business processes, supervision (video & crowd) operators’ training, data generation for deep neural networks training (video & dataset generation), decision support, what-if scenarii exploration and collaborative systems testbed.

SE-Star can provide information from high level (density, alarms, situation reports) to low level (individual’s information, devices’ status)

SE-Star can simulate between 5000 and 10000 peoples on one computer. SE-Star can work in distributed mode pushing limits. (Tested with 80000 peoples on 10 NUC computers)

SE-Star has been used in EU project (Opti-Alert, SECURE-Ed, iCore) and in THALES projects covering airports (Pisa), train stations (Gare du Nord) or crowded places (Mecca).

Easily extendable, SE-Star has been connected to real systems (crowd monitoring, airport supervision) using standard interfaces and protocols (SAOP, REST, RTSP)

See SE-Star in action : http://www.youtube.com/watch?v=cTM34ReycqM&t=2m12s

Readiness
Crisis size
Crisis Cycle Phase
SE-Star : THALES Crowd Simulation - Gare du Nord
SE-Star : THALES Crowd Simulation - Parc des Princes
SE-Star : THALES Crowd Simulation - New-York