Crowd simulation for digital twin applications

stock-photo-pedestrian-blur-crowd-of-people-walking-in-london-city-panoramic-view-of-people-crossing-the-2316302269-1x1

Crowd Simulation for Digital Twin Applications

This project aims to advance crowd simulation functionality for Digital Twin purposes, through the adaptation of Agent-based crowd simulation up to extreme densities using fluid dynamics.

Jonathan Lam

Master Student, NUS Computing

stock-photo-pedestrian-blur-crowd-of-people-walking-in-london-city-panoramic-view-of-people-crossing-the-2316302269

Campus as a Living Lab (CALL@NUS)

Crowd Simulation for Digital Twin Applications

This project aims to advance crowd simulation functionality for Digital Twin purposes, through the adaptation of Agent-based crowd simulation up to extreme densities using fluid dynamics.

Jonathan Lam

Master Student, NUS Computing

This project worked towards the implementation of realistic crowd simulation capability for large-scale crowds.

It adapted the method found in SPH crowds: Agent-based crowd simulation up to extreme densities using fluid dynamics (van Toll et al, 2021) in order to produce stable density-controllable crowds.

Integrating and adding to the method found in Exact wavefront propagation for globally optimal one-to-all path planning on 2d cartesian grids (Ibrahim et al, 2024), it further provided path-planning and dynamic exit switching capabilities, that are tractable at scale, to the simulation system.

Finally, it also attempted to visualize crowd simulations meaningfully in the form of 2D heatmaps and the 3D Digital Twin set-up using Unity. Further work could include in-depth crowd analytics, user-friendly UI, and deeper integration with the current NUS Digital Twin set-up.

About the author

I am a 5th Year Computer Engineering and Economics DDP Student with interests in Computer Graphics and Artificial Intelligence.

The research on crowd simulation specifically for digital twin applications held special appeal to me for the wide range of fields it spans.

Through it, I was able to expose myself to machine learning methods for crowd simulation, algorithmic optimizations, physics simulations, and cloud deployment.

Above all, I cherish the freedom afforded as a research assistant on the NUS Digital Twin Project to challenge myself against existing and new problems and grow both as a researcher and as an engineer.

More about CALL@NUS

Sidebar

Main Menu

About the author