Modeling of traffic flow of automated vehicles

With the development of near term automatic vehicles following concepts such as intelligent cruise control (ICC) and cooperative driving, vehicles will be able to automatically follow each other in the longitudinal direction. The modeling of traffic flow consisting of such vehicles is important for analyzing the effects of vehicle automation on the characteristics of traffic flow and for suggesting macroscopic control strategies to improve efficiency. Such analysis may also suggest ways for modifying the vehicle control characteristics in order to improve the macroscopic behavior of traffic. In this paper, we developed a mesoscopic and macroscopic model that describes the automated traffic-flow dynamics in a single highway lane. The mesoscopic model describes the speed and density continuously in time and space and at the same time retains the microscopic characteristics of traffic flow. The macroscopic model describes the average speed and density at each section of the lane and at each point in time. Even though the macroscopic model does not retain the microscopic characteristics of the vehicular traffic, computationally it is much simpler than the mesoscopic one. Simulations are used to demonstrate the effectiveness of these models in describing traffic-flow characteristics. The developed models indicate some similarities, but also some fundamental differences with existing traffic-flow models for manually driven vehicles.

I. Introduction

In order to increase the capacity and safety of the existing highway system, a variety of concepts have been introduced that allow vehicles to follow each other automatically with varying degree of authority between human drivers and automatic control laws. Some of these concepts, which have the potential to appear in the near future without extensive infrastructure modifications, are the intelligent cruise control (ICC) and cooperative driving systems. Vehicles equipped with ICC can automatically follow other automated or manually driven vehicles in the same lane; hence, they can provide partial or even full automation in the longitudinal direction. On the other hand, cooperative driving systems require some exchange of information between neighboring vehicles and/or infrastructure to coordinate maneuvers and hence may require a dedicated lane where only automated vehicles are allowed to operate. As the number of vehicles with automatic vehicle-following features increases, the behavior of traffic flow will change in a way that needs to be understood.