Macroscopic Dynamics of Multi-Lane Traffic

TL;DR

This paper introduces a macroscopic, empirically calibrated model for multi-lane freeway traffic that captures lane interactions, lane-changing dynamics, and various traffic phenomena through non-local, anisotropic equations.

Contribution

It presents a novel gas-kinetic based macroscopic model for multi-lane traffic, including lane-changing rates and validated with Dutch highway data.

Findings

Model accurately reproduces empirical traffic data

Enables simulation of lane synchronization and disruptions

Facilitates analysis of traffic management strategies

Abstract

We present a macroscopic model of mixed multi-lane freeway traffic that can be easily calibrated to empirical traffic data, as is shown for Dutch highway data. The model is derived from a gas-kinetic level of description, including effects of vehicular space requirements and velocity correlations between successive vehicles. We also give a derivation of the lane-changing rates. The resulting dynamic velocity equations contain non-local and anisotropic interaction terms which allow a robust and efficient numerical simulation of multi-lane traffic. As demonstrated by various examples, this facilitates the investigation of synchronization patterns among lanes and effects of on-ramps, off-ramps, lane closures, or accidents.