Challenges in Crowd Simulation

  title={Challenges in Crowd Simulation},
  author={Daniel Thalmann and Helena Grillon and Jonathan Ma{\"i}m and Barbara Yersin},
  journal={2009 International Conference on CyberWorlds},
The purpose of this paper is to identify the problems to solve in order to simulate real-time crowds in a Virtual Environment. We try to classify these problems and study how they have been addressed until now by the research community and our Lab in particular. We then discuss for each problem what are the future challenges and how to address them. 

Figures from this paper

A Hybrid Model for Simulating Crowd with Varied Behaviors in Real-Time

A model based on a macroscopic model is proposed, which, mainly, aims at reducing the cost of computing the potential fields of the groups, by using groups discretized according to the needs of the environment.

Modelling Crowds in Urban Spaces

This work uses crowd sourced location data to compute spatio-temporal people and vehicle flows, while map and geometric data describe specific real places to visualize both real trajectories and data driven on-line crowd simulation.

Fuzzy and Data-Driven Urban Crowds

This work presents a system able to simulate crowds in complex urban environments by incorporating steering pedestrian behaviors with the goal of simulating the subtle variations present in real scenarios without needing large amounts of data for those low-level behaviors.

Crowd simulation and visualization

This paper presents a methodology to simulate and visualize crowds, using real data from which the behavior of the masses is identified applying statistical and artificial intelligence techniques.

Uniform spatial subdivision to improve Boids Algorithm in a gaming environment

This document implements and test one of these performance optimization methods, and uses benchmarking results to compare the performance of the method versus the Boids Algorithms’ brute force neighborhood gathering approach.

Research on Crowd Behavior Based on the Velocity Field

Data-driven modeling method has become an ideal research method in crowd behavior simulation fields. The basic idea of data-driven modeling method is that velocity field generated by the crowd

Proximity Queries for Crowd Simulation Using Truncated Voronoi Diagrams

A parallel technique, using consumer grade graphics hardware, for proximity queries that is suitable for real-time crowd simulations, and has significantly better performance than similar methods and can achieve simulations with thousands of agents in interactive frame rate.

Shadow obstacle model for realistic corner-turning behavior in crowd simulation

This paper describes a novel model known as the shadow obstacle model to generate a realistic corner-turning behavior in crowd simulation and performs parameter analysis to show the believability of the model through a series of experiments.



Continuum crowds

In this model, a dynamic potential field simultaneously integrates global navigation with moving obstacles such as other people, efficiently solving for the motion of large crowds without the need for explicit collision avoidance.

Crowd patches: populating large-scale virtual environments for real-time applications

This paper extends the concept of motion patches to densely populate large environments and provides a drastic lowering of computation needs for simulating a virtual crowd at run-time, which can handle dense populations in large-scale environments with performances never reached so far.

Unique Character Instances for Crowds

A proposed solution varies the shape of human instances by attaching accessories and modifies the instances' appearance with a generic technique based on segmentation maps that can generate detailed color variety and patterns.

Hierarchical model for real time simulation of virtual human crowds

  • S. MusseD. Thalmann
  • Computer Science
    5th IEEE EMBS International Summer School on Biomedical Imaging, 2002.
  • 2002
A model for simulating crowds of humans in real time composed of virtual crowds, groups and individuals with the possibility of increasing the complexity of group/agent behaviors according to the problem to be simulated, and the hierarchical structure based on groups to compose a crowd.

Unique Instances for Crowds

Afficher la publication dans Web of Science Notice creee le 2008-06-25, modifiee le 2012-07-10.

Crowds of Moving Objects: Navigation Planning and Simulation

This paper presents a solution to interactive navigation planning and real-time simulation of a very large number of entities moving in a virtual environment by deducing a structure called navigation graph from the environment geometry analysis, which is the base to the method.

Clone attack! Perception of crowd variety

In a series of experiments designed to test people's perception of variety in crowds, it is found that clones of appearance are far easier to detect than motion clones and that cloned models can be masked by color variation, random orientation, and motion.

Real‐time navigating crowds: scalable simulation and rendering

This paper introduces a framework for real‐time simulation and rendering of crowds navigating in a virtual environment using a specific environment preprocessing technique giving rise to navigation graphs, which are then used by the navigation and simulation tasks.

Visualizing Crowds in Real‐Time

Methods to deal with various aspects of crowd visualization, ranging from collision detection and behaviour modeling to fast rendering with shadows and quality shading are looked at, to suggest that simulations of reasonably complex environments populated with thousands of animated characters are possible in real‐time.

Real-Time Scalable Motion Planning for Crowds

A hybrid architecture to handle the path planning of thousands of pedestrians in real time, while ensuring dynamic collision avoidance and ensuring pedestrian motion continuity when switching between motion planning algorithms is presented.