Pedestrian Dynamics - Modeling


Our research is guided by the following research question: How can we understand the properties of a crowd and predict its dynamics in closed and open spaces?

This question is addressed through a research design that weaves together three streams: development of operative models, investigation of route choice models, and software engineering. In this way, our division aims for interdisciplinary research.

In the long-term different products for prediction and evaluation of crowds dynamics are to be developed based on different open-source projects to guarantee sustainability and continuation of development. In this regard, our research is centered around the development of JuPedSim, an open-source software mainly developed by researchers from our division. It will serve as starting platform for investigating new scientific questions related to modeling of pedestrian dynamics as well as developing new solutions for real-time assessment and visualization of crowd densities in buildings and open spaces.

Simulation of pedestrians exiting a complex underground station

Our research is used in different ways. First, bachelor, master, and Ph.D. students are provided with numerical frameworks to focus on research, i.e. development and validation of new models. Their research results are not only published in written form but will be made public in the form of software pieces integrated into the offered framework. In this manner, we can guarantee continuity of research and strengthen the reproducibility of scientific results even after the end of the supervised theses.
Second, we aim to bundle our theoretical findings in form of a real-time assessment and prediction tool that can be used in different applied projects to gain instant insights on the dynamics of the crowd and hence preventively avoid potentially dangerous situations e.g. high densities as well as temporal and spatial bottlenecks.

The research of our division aims at exploring several interrelated topics that fall under "mathematical modeling of pedestrian dynamics". These topics are:

  • Identification and analysis of both, micro and macro factors, that affect pedestrian movement in real-world application scenarios. Our goal in this context is to quantify and measure the level of impact of each factor on pedestrians' movement decisions and behavior.
Simulation of a lane-formation phenomenon: lanes automatically occur in bi-directional flows as a direct result of minimizing conflicts in different streams.
  • Exploiting a variety of modeling techniques to replicate and simulate crowd movement features, reaching from physical-driven modeling designed to understand key characteristics of crowd movement to data-driven models aiming at predicting the outcome of evacuation processes. For instance, analyzing crucial bottlenecks in evacuation scenarios.
  • Investigation of many route choice strategies and their influence on the overall evacuation time from a building.
Pedestrians are leaving the building with different routing strategies. This simulation allows to investigate the influence of the proportion of agents deciding to evacuate through a certain exit.
Evacuation time concerning the proportion of evacuees from a certain exit show a certain optimal minimum.

Research Topics

Last Modified: 14.06.2022