Skip directly to content

Crowd Control Based on Individuals

Two groups of researchers have created computer models that assess the behavior of individuals in large groups. These models are particularly powerful because they integrate individual characteristics such as age, ethnicity and gender, to predict behavior at the group level.

The models will be a valuable tool for police or military in planning and handling densely populated events, especially in emergency situations. The predictive and experimental power of the models can also assist in urban design and city planning.

The model, developed by a team based initially at Arizona State University and now at the University of Maryland, integrates census data with motion-capture data to build a fully immersive 3-D environment. The environment is reusable and behaviorally founded.

To create a model of individual movement in social situations, a group at the University of Central Florida filmed crowds in many scenarios such as leaving church, changing class at school and leaving a football game. The diverse demographic make-up of the crowds on film allow the researchers to build models of pedestrians that incorporate demographic as well as personality data. Researchers can use these types of models to compare and predict crowd behavior in panicked or calm scenarios and to design more effective evacuation strategies.


  • computer created crowd scene
Computer models explore crowd dynamics.
Paul M. Torrens

Recent Award Highlights

a researcher works with a !xuun speaker in namibia

Rare speech sounds discovered in African languages

Unusual click consonants in languages of Botswana and Namibia

Research Areas: People & Society Locations: Arizona New York International
a section of the dubulay archeological site in guyana

The impact of early agricultural changes in Guyana

Findings suggest sustained human settlements over multiple millenia

Research Areas: People & Society Locations: Florida International