Discovering interaction mechanisms in crowds via deep generative surrogate experiments

Abstract Understanding pedestrian crowd dynamics is a fundamental challenge in active matter physics and crucial for efficient urban infrastructure design. Complexity emerges from social interactions, which are often qualitatively modeled as distance-based additive forces. Endeavors towards quantita...

Full description

Saved in:
Bibliographic Details
Main Authors: Koen Minartz, Fleur Hendriks, Simon Martinus Koop, Alessandro Corbetta, Vlado Menkovski
Format: Article
Language:English
Published: Nature Portfolio 2025-03-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-92566-9
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850208549548851200
author Koen Minartz
Fleur Hendriks
Simon Martinus Koop
Alessandro Corbetta
Vlado Menkovski
author_facet Koen Minartz
Fleur Hendriks
Simon Martinus Koop
Alessandro Corbetta
Vlado Menkovski
author_sort Koen Minartz
collection DOAJ
description Abstract Understanding pedestrian crowd dynamics is a fundamental challenge in active matter physics and crucial for efficient urban infrastructure design. Complexity emerges from social interactions, which are often qualitatively modeled as distance-based additive forces. Endeavors towards quantitative characterizations have been limited by a trade-off between parametric control in laboratory studies and statistical resolution of large-scale real-world measurements. To bridge this gap, we propose a virtual surrogate experimentation paradigm that combines laboratory-like control with real-world statistical resolution. Our approach hinges on a generative simulation model based on graph neural networks, which we train on real-world pedestrian tracking data and validate against key statistical properties of crowd dynamics. Our surrogate experiments not only reproduce known experimental results on collision avoidance, but also reveal new insights into N-body interactions in crowds, which have remained poorly understood. We find that these interactions are topological, with individuals reacting to a limited number of neighbors within a narrow field of view. Our study exemplifies how data-driven approaches can uncover fundamental interaction structures in social systems, even when only uncontrolled measurements are available. This approach opens new avenues for scientific discovery in complex systems where laboratory studies are prohibitive, from crowd dynamics and animal behavior to opinion formation.
format Article
id doaj-art-5ea9e21c9d3441e48a08a9cd0d005f3a
institution OA Journals
issn 2045-2322
language English
publishDate 2025-03-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-5ea9e21c9d3441e48a08a9cd0d005f3a2025-08-20T02:10:13ZengNature PortfolioScientific Reports2045-23222025-03-0115111110.1038/s41598-025-92566-9Discovering interaction mechanisms in crowds via deep generative surrogate experimentsKoen Minartz0Fleur Hendriks1Simon Martinus Koop2Alessandro Corbetta3Vlado Menkovski4Department of Mathematics and Computer Science, Eindhoven University of TechnologyDepartment of Mechanical Engineering, Eindhoven University of TechnologyDepartment of Mathematics and Computer Science, Eindhoven University of TechnologyDepartment of Applied Physics and Science Education, Eindhoven University of TechnologyDepartment of Mathematics and Computer Science, Eindhoven University of TechnologyAbstract Understanding pedestrian crowd dynamics is a fundamental challenge in active matter physics and crucial for efficient urban infrastructure design. Complexity emerges from social interactions, which are often qualitatively modeled as distance-based additive forces. Endeavors towards quantitative characterizations have been limited by a trade-off between parametric control in laboratory studies and statistical resolution of large-scale real-world measurements. To bridge this gap, we propose a virtual surrogate experimentation paradigm that combines laboratory-like control with real-world statistical resolution. Our approach hinges on a generative simulation model based on graph neural networks, which we train on real-world pedestrian tracking data and validate against key statistical properties of crowd dynamics. Our surrogate experiments not only reproduce known experimental results on collision avoidance, but also reveal new insights into N-body interactions in crowds, which have remained poorly understood. We find that these interactions are topological, with individuals reacting to a limited number of neighbors within a narrow field of view. Our study exemplifies how data-driven approaches can uncover fundamental interaction structures in social systems, even when only uncontrolled measurements are available. This approach opens new avenues for scientific discovery in complex systems where laboratory studies are prohibitive, from crowd dynamics and animal behavior to opinion formation.https://doi.org/10.1038/s41598-025-92566-9Crowd dynamicsPedestrian dynamicsActive matter physicsGenerative modelsGraph neural networksNeural simulators
spellingShingle Koen Minartz
Fleur Hendriks
Simon Martinus Koop
Alessandro Corbetta
Vlado Menkovski
Discovering interaction mechanisms in crowds via deep generative surrogate experiments
Scientific Reports
Crowd dynamics
Pedestrian dynamics
Active matter physics
Generative models
Graph neural networks
Neural simulators
title Discovering interaction mechanisms in crowds via deep generative surrogate experiments
title_full Discovering interaction mechanisms in crowds via deep generative surrogate experiments
title_fullStr Discovering interaction mechanisms in crowds via deep generative surrogate experiments
title_full_unstemmed Discovering interaction mechanisms in crowds via deep generative surrogate experiments
title_short Discovering interaction mechanisms in crowds via deep generative surrogate experiments
title_sort discovering interaction mechanisms in crowds via deep generative surrogate experiments
topic Crowd dynamics
Pedestrian dynamics
Active matter physics
Generative models
Graph neural networks
Neural simulators
url https://doi.org/10.1038/s41598-025-92566-9
work_keys_str_mv AT koenminartz discoveringinteractionmechanismsincrowdsviadeepgenerativesurrogateexperiments
AT fleurhendriks discoveringinteractionmechanismsincrowdsviadeepgenerativesurrogateexperiments
AT simonmartinuskoop discoveringinteractionmechanismsincrowdsviadeepgenerativesurrogateexperiments
AT alessandrocorbetta discoveringinteractionmechanismsincrowdsviadeepgenerativesurrogateexperiments
AT vladomenkovski discoveringinteractionmechanismsincrowdsviadeepgenerativesurrogateexperiments