Getting Out of the Way: Collision-Avoiding Pedestrian Models Compared to the RealWorld

Abstract

AbstractNumerical simulation of human crowds is a challenging task and a number of models to simulate pedestrian dynamics on a microscopic level have been established. One aim of those models is to reproduce a realistic, and in particular collision-free, movement of crowds in complex environments. This work investigates three approaches on their capability to reproduce a collision-free movement of pedestrian crowds in complex dynamic environments. The baseline model is the well-known social force model. While in the social force model pedestrians do not explicitly avoid each other, the second model extends the social force model to avoid collisions explicitly. The observed collision-avoiding behavior produced by the third model is reached by calculating velocity obstacles. These are obstacles in the velocity space, meaning that if a pedestrian chooses a velocity that lies inside the velocity obstacle, then a collision occur at some time. This work discusses the models and their integration in a multi-agent simulation framework. The models are tested on data from a real-world experiment conducted at Technische Universität Berlin. In this experiment, two pedestrian flows intersected at an angle of 90∘. The models’ performance with regard to the reproduction of a realistic crowds movement and their computational complexity are discussed in this work.KeywordsIntersecting pedestrian flowsMulti-agent simulationHuman crowd experiments