Social Force based Vehicle Model for Two-Dimensional Spaces

Abstract

There is a growing interest in modeling heterogeneous traffic. This is especially true in Asian countries where a wide variety of transport modes operate on streets under differing traffic rules. There is also growing interest by urban planners and traffic engineers in applying the shared space or living street concepts to roadways. These approaches remove the traditional segregation of different road users by replacing lane markings and curbs with open spaces for use by all road users. There is some debate including safety and capacity issues when integrating pedestrian, bicyclists and car movement on an open space with little guidance. Current microscopic traffic flow models are not suitable when pedestrians, bicycles, cars and heavy vehicles share the same space since these modes have very different levels of maneuverability and speeds. Most current microscopic traffic flow models are 1- dimensional; they are based on car-following modes with limited consideration of lateral interaction. In contrast, social force models, frequently used to model pedestrian behavior, consider 2-dimensional motion and are not link-based. These models use an interacting force between pedestrians and obstacles to model walking and are often applied for crowd modeling. This paper presents an approach for extending the social force model to models also for vehicles movements. In contrast to traditional car-following models, the proposed approach uses a mechanically-based model to obtain reasonable turning trajectories and a proportional–integral–derivative (PID) controller is integrated to control the simulated vehicles. The method was tested using data collected by video imaging at a T-intersection that was recently converted to a shared space area. Results of this case study show that the trajectories simulated using the proposed model are good approximations of the real trajectories.