Description
Pedestrian dynamics simulation is the numerical execution of any crowd model — social-force-model, cellular-automata, continuum-crowd-model, or hybrids — to produce trajectories, density fields, or flow fields under a given scenario. In the thesis it is the toolchain that produces synthetic ground truth (BLE-style trajectories + CSI-style perturbation forecasts) for evaluating the BLE-calibrated CSI pipeline before deployment.
When it's used
- Generating synthetic ground truth for new sensing methods
- Capacity planning for evacuation / event scenarios
- Sensitivity analyses (how does CSI response change with crowd parameters?)
- Counterfactual reasoning when only sparse real data is available
Limitations
- Validation against real crowds is the hard part, not the simulation itself
- Different modelling families produce qualitatively different macroscopic behaviour
- Computational cost scales poorly for very large agent counts