Model-based assessment of the impact of driver-assist vehicles using kinetic theory

TL;DR

This paper uses kinetic theory to analyze how driver-assist strategies influence traffic flow at different scales, revealing that local effects do not necessarily translate to macroscopic traffic improvements.

Contribution

It provides a novel kinetic framework to evaluate the multiscale impact of driver-assist controls on traffic dynamics, highlighting their limited effect on overall traffic throughput.

Findings

Control strategies affect local traffic features like speed and headway dispersion.

Such strategies have minimal impact on macroscopic traffic flow metrics.

The kinetic approach aligns with recent field observations on autonomous vehicles.

Abstract

In this paper we consider a kinetic description of follow-the-leader traffic models, which we use to study the effect of vehicle-wise driver-assist control strategies at various scales, from that of the local traffic up to that of the macroscopic stream of vehicles. We provide a theoretical evidence of the fact that some typical control strategies, such as the alignment of the speeds and the optimisation of the time headways, impact on the local traffic features (for instance, the speed and headway dispersion responsible for local traffic instabilities) but have virtually no effect on the observable macroscopic traffic trends (for instance, the flux/throughput of vehicles). This unobvious conclusion, which is in very nice agreement with recent field studies on autonomous vehicles, suggests that the kinetic approach may be a valid tool for an organic multiscale investigation and possibly design of driver-assist algorithms.