A Simple Model for Phase Separation in Pedestrian Dynamics

Abstract

The occurrence of phase separation is a common feature observed in vehicular traffic. Experiments have shown a similar behavior for pedestrians, though the situation in pedestrian dynamics is more complicated. The two separate phases in one-dimensional “single-file” pedestrian movement are a jammed high-density phase and a phase of medium to high density with slowly moving pedestrians. Both phases consist of interacting particles (pedestrians). To understand the emergence of this kind of phase separation, we develop a simple cellular automaton model. The transition probabilities of the modeled pedestrians in general depend on their current velocities and on the occupancy of the next two cells in front of them. For inhomogeneous initial conditions the simulated pedestrian trajectories clearly feature two distinct phases analogous to the experiment. The lifetime of the decaying jammed phase is consistent with experimental results.