Modular Modelling for Urban Traffic Networks based on Multi-Agent Systems and Petri Nets

Inspired on the focus of multi-agent systems used in groups of robots, this paper describes a novel modeling methodology for urban traffic networks based on the interconnections of simple generic differential equations about distinct streets and intersections. These nonlinear fluid equations represent features like vehicular density, average vehicle velocities, occupation capacities and the interconnections between adjacent street segments and intersections. In order to enable more realistic traffic light actions, individual petri nets are defined for each the intersections, enabling or disabling the flow of vehicles. Therefore, the main contribution is the definition of a hybrid modular scheme of continuous differential equations with discrete-event systems. A numerical simulation of a simple intersection using Matlab shows the performance of the approach and its scalability to more complex traffic networks.