CrowdPetri

CrowdPetri is a web + desktop app for agent-based modeling of pedestrian crowds in venues (stadiums, transit hubs, festivals, campuses). Users import a floorplan (image/PDF), draw walkable areas, entrances/exits, and obstacles, then define agent types (commuters, families, staff, mobility-impaired) with speed, patience, and route-choice behaviors. The simulator runs scenarios like partial gate closures, delayed trains, security checks, weather constraints, or emergency egress, producing heatmaps, density-over-time charts, queue lengths, and “near-crush” risk indicators. It’s a combination traditional + AI app: AI assists with auto-tracing floorplans into navigable meshes and suggests parameter defaults from past runs, but the core model remains transparent and editable. Export reports for safety reviews and operations planning. Realistically, accuracy depends heavily on good calibration data; without it, outputs are directional rather than definitive.