Title:High-statistics modeling of complex pedestrian avoidance scenarios

Abstract:Quantitatively modeling the trajectories and behavior of pedestrians walking in crowds is an outstanding fundamental challenge deeply connected with the physics of flowing active matter, from a scientific point of view, and having societal applications entailing individual safety and comfort, from an application perspective.
In this contribution, we review a pedestrian dynamics modeling approach, previously proposed by the authors, aimed at reproducing some of the statistical features of pedestrian motion. Comparing with high-statistics pedestrian dynamics measurements collected in real-life conditions (from hundreds of thousands to millions of trajectories), we modeled quantitatively the statistical features of the undisturbed motion (i.e. in absence of interactions with other pedestrians) as well as the avoidance dynamics triggered by a pedestrian incoming in the opposite direction. This was accomplished through (coupled) Langevin equations with potentials including multiple preferred velocity states and preferred paths. In this chapter we review this model, discussing some of its limitations, in view of its extension toward a more complex case: the avoidance dynamics of a single pedestrian walking through a crowd that is moving in the opposite direction. We analyze some of the challenges connected to this case and present extensions to the model capable of reproducing some features of the motion.