Emergent Crossing Regimes of Identical Autonomous Vehicles at an Uncontrolled Intersection

TL;DR

This paper explores how identical autonomous vehicles behave at uncontrolled intersections across different traffic densities, revealing distinct crossing regimes and analytical predictions of transitions, aiding development of better AV decision-making.

Contribution

It introduces a simple MATLAB simulation to identify fundamental crossing regimes and analytically predict transitions, advancing understanding of AV behavior at intersections.

Findings

Distinct crossing regimes identified at different densities

Transitions between regimes can be predicted analytically

Performance linked to fundamental 1-D traffic flow model

Abstract

To investigate the impact of Autonomous Vehicles (AVs) on urban congestion, this study looks at their performance at road intersections. Intersection performance has been studied across a range of traffic densities using a simple MATLAB simulation of two intersecting 1-D flows of homogeneous automated vehicles. This lacks the detail of more advanced simulations, but it enables fast identification of fundamental behaviours. The results show that there are distinct crossing regimes at low, medium and high densities. Furthermore, the transitions between regimes can be predicted analytically and their performance related to the fundamental model of 1-D traffic flow. These findings have the potential to focus efforts on the development of improved decision-making rules for emerging AVs.