Navigation and service

ICS-2 Talk: A cognitive, decision-based model for pedestrian dynamics

Cornelia von Krüchten, Universität zu Köln

begin
12 Dec 2019 11:00
venue
Room 2001, Building 04.16

Abstract:

In the light of growing cities and mass events, understanding and analysing pedestrian behaviour is crucial. From a physicist’s point of view, pedestrian crowds can be understood as a system of active particles that additionally have psychological and sociological mechanisms. As experiments on pedestrian behaviour can be difficult or even impossible for safety or logistical reasons, research on pedestrian dynamics also heavily relies on model development and simulations.

In this talk a new model approach is presented which aims at capturing a pedestrian’s decision process while walking. Based on information on the environment and other pedestrians an agent determines his/her velocity for the next time step. Including perception, navigation or anticipation, the model considers cognitive abilities explicitly. Further underlying psychological mechanisms are represented by stochastic elements which model inherent uncertainties of human decisions.

The model is validated and calibrated in comparison to experimental data. Pedestrian crowds exhibit interesting collective phenomena like stop-and-go waves in single-file motion or lane formation in counterflow. It can be shown that the model is able to represent collective phenomena for several characteristic scenarios of pedestrian dynamics. The respective parameter adjustments help at gaining insight into the decision processes of pedestrians.


Contact

PD Dr. Jens Elgeti
Phone: +49 2461 61-9382
Fax: +49 2461 61-3180
email: j.elgeti@fz-juelich.de