Agent-based crowd simulation is a computational modeling approach in which individual pedestrians or entities are represented as autonomous agents that follow defined behavioral rules — such as force-based dynamics or cellular automata transition rules — to collectively reproduce emergent crowd movement patterns at a microscopic level. It matters for WiFi/CSI sensing and crowd analytics research because it provides a principled framework for generating synthetic ground-truth data, validating sensing-derived crowd estimates, and enabling real-time forecasting of crowd flows when direct observation is incomplete or impractical. Key variants include force-based models, which simulate physical and social forces acting on each agent, and data-driven cellular automata models, which discretize space into adjustable-resolution grids and learn transition rules from empirical data; both can be coupled with data assimilation techniques such as particle filters to continuously update simulations with aggregate real-world observations.
Source Papers
- A crowd team evacuation model considering spring effect ↗ — A crowd team evacuation model considering spring effect
- Crowd flow forecasting via agent-based simulations with sequential latent parameter estimation from aggregate observation ↗ — Crowd flow forecasting via agent-based simulations with sequ
- Modeling spatial patterns in a moving crowd of people using data-driven approach—A concept of Interplay Floor Field ↗ — Modeling spatial patterns in a moving crowd of people using
- Recent trends in crowd analysis: A review ↗ — Recent trends in crowd analysis: A review
- State-of-the-art crowd motion simulation models ↗ — State-of-the-art crowd motion simulation models