Simulation of Pedestrian Flow at a Crosswalk by a Multi-Agent System with a Pre-Avoid Forces Model

The problem of pedestrian crossings is one of the topics that traditionally have played an important role in pedestrian dynamics research. A pedestrian flow dynamics simulation will lead to a system that serves as a rational tool for the urban pedestrian crossing design process to better understand how facilities designs will influence pedestrian behavior. The article describes a multi-agent system (MAS) method with a social force model in pedestrian dynamics modeling at a crossing, whose primary objective is to develop a microscopic dynamics model for a bi-directional pedestrian flow at a crossing. In this paper, we newly develop a pre-avoid forces model, which can describe either a collision avoidance or the lane formation phenomenon between adjacent pedestrians. The MAS method with the pre-avoid forces model shows good predictions of the characteristics to bi-directional pedestrian flow at a crossing.