A two-level urban traffic control for autonomous vehicles to improve network-wide performance

In the near future, autonomous vehicles will face with new challenges in several fields. One of the most exciting changes will be represented by the network-wide optimal traffic control. When driverless vehicles take over the road, classical road signalization schemes will become superfluous. Accordingly, the paper’s aim is to propose a control design methodology for autonomous vehicles in urban traffic network by considering the network-wide performance. The proposed two-level control strategy solves a tractable optimization problem for a network wide traffic control. On the one hand, a local intersection controller is designed which ensures safe crossings of vehicles and aims to reduce traffic emission in the junction area. On the other hand, the local controllers also optimize the network performance by minimizing the queues in all road links. The traffic is therefore modeled in a two-level fashion. A microscopic dynamics is considered in junctions and macroscopic model is applied for the whole traffic network. The control strategy is tested and evaluated based on microscopic traffic simulation.