TY  - CONF
AU  - Seyfried, Armin
TI  - Congestion in crowds – behaviour and risks at high densities and in crowds
PB  - George Washington University
M1  - FZJ-2025-05782
PY  - 2025
AB  - Predicting congestion in pedestrian flows is useful for planning events, transportation hubs, or escape routes in buildings. According to the state of the art, software solutions are based on agent-based models in which pedestrians are represented as two-dimensional objects (e.g., circles, ellipses, etc.). These models are able to predict congestion in complex path networks, but reach their limits when it comes to describing crowds at high densities.The talk will use witness statements from the Love Parade in Duisburg (an event in which 31 people died) to analyse how people behave in crowds and which dynamics lead to life-threatening situations. In a second part, a methodology will be presented for collecting data that provides a three-dimensional description of the movement and interaction of bodies (torsos and limbs) in crowds. This data is used to develop hybrid AI models in which pedestrians interact as three-dimensional objects. Using methods and concepts from social psychology, we are working on models in which the dynamic of motivation changes is described, or which analyse the spread of behaviour in crowds. Empirical data from laboratory experiments confirm the findings from the analysis of witness statements and show how slowly pushing and shoving behaviour spreads in crowds.
T2  - Transportation Engineering Seminar Series
CY  - , Washington D.C. (USA)
M2  - Washington D.C., USA
LB  - PUB:(DE-HGF)31
UR  - https://juser.fz-juelich.de/record/1050073
ER  -