A Data-Driven Integrated Framework for Virtual Testing of Autonomous Vehicles in Mixed Traffic Scenarios

The increasing presence of autonomous vehicles (AVs) in urban environments introduces both opportunities and challenges, particularly regarding their interactions with traditional vehicles and other road users. This paper presents a comprehensive framework designed to assess the integration of AVs in mixed traffic scenarios. The framework is built upon real-world data collected from AV trials conducted in Turin, Italy. By leveraging traffic microsimulation along with machine learning techniques, the study proposes a framework aimed at assessing ex-ante the impacts on traffic of AVs introduction, thus constituting a relevant tool of virtual testing of CCAM (Cooperative, Connected, and Automated Mobility) trials before the physical introduction of autonomous vehicles on public roads. The integration of High-Performance Computing (HPC) ensures the efficiency of these simulations, enabling real-time analysis and testing. The proposed framework not only provides decision-makers with a tool for virtual testing of AV deployment, but also offers actionable insights into traffic management strategies. The study’s findings contribute to a deeper understanding of the role AVs can play in future urban mobility systems, particularly as cities prepare for the broader adoption of CCAM technologies.