Fine-Grain Traffic Control for Smart Intersections

Abstract

As connected and, even more, autonomous vehicles are expected to bring significant novelties in the future road traffic patterns, we have investigated the control of a specific, yet very common topology, such as the intersection between two 2-lane roads. We have addressed the issue with a novel, fine-grain control approach, and proposed an adaptive prioritization algorithm which weights length of the queue and arrival order for each lane. From an Uppaal simulation, we deduce that the second factor looks more important, at higher arrival rates. Compared to a fixed Round-robin schedule, our algorithm achieves quite a better performance, especially at high traffic volumes, also with inhomogeneous traffic flow cases. In order to guarantee robustness to our design, we made a model checking analysis, considering safety and liveness requirements.