Generic Multi-class Cell Transmission Model for Traffic Control

Mladen Cicic, Enrico Siri, Maria Laura Delle Monache

Abstract: Recent years have witnessed renewed interest in multi-class traffic models, inspired in no small part by the impending arrival of Connected and Autonomous Vehicles, whose behaviour is likely to differ from that of Human-Driven Vehicles. Although numerous multi-class traffic models have been proposed, consistent overarching theory is lacking. In this paper, we propose a generic first-order multi-class traffic modelling framework, intended to be sufficiently versatile to represent most of the traffic phenomena relevant to freeway control applications. Based on this framework, we are able to instantiate different specific multi-class models by appropriate choice of a few design functions. We restrict the design space by introducing a set of assumptions that these functions should follow, helping guide the modelling process. Finally, we study a simple control example where a single class of vehicles is controlled in order to dissipate congestion on a highway, and test the control law using several multi-class model variants. The simulation results show that proposed control is able to dissipate the congestion and harmonize the traffic without relying on knowing the exact underlying traffic dynamics.