News

The student Joseba Gorospe Jauregui obtained an EXCELLENT grade with mention International Doctorate

---

• Thesis title: Resilient Cooperative Automated Driving

Court:

• Presidency: Ramiro Sámano Robles (Instituto Superior De Engenharia Do Porto)
• Vocal: Michele Segata (University of Trento)
• Secretary: Maitane Barrenechea Carrasco (Mondragon Unibertsitatea)

Abstract:

Vehicle-to-Everything (V2X) communications are critical to enable Autonomous Vehicles (AVs) to interact and coordinate with each other. Their continuous information exchange facilitates cooperative driving maneuvers that improve traffic safety and efficiency. Despite these benefits, V2X communication technologies face significant challenges, particularly in terms of their reliability in some scenarios. This thesis investigates the impact of this unreliability during cooperative driving maneuvers and explores the potential for improving control algorithms by considering low-reliability scenarios, as well as incorporating Machine Learning (ML) based predictions into the system. Initially, field tests are conducted to evaluate the performance of existing V2X communication technologies, such as IEEE 802.11p and LTE-V2X. These tests are replicated and extended in a simulation framework to compare the technologies in more complex scenarios involving cooperative driving maneuvers such as platooning or emergency braking. These comparisons are used to identify the most suitable technology for safety related cooperative driving maneuvers. Then, some ML models with a data-driven approach have been used to anticipate hazardous scenarios where a platoon may struggle to avoid inter-vehicle collisions during emergency braking. These ML models consider both platoon control parameters and contextual information related to V2X communication channel, such as traffic density or message transmission intervals. Although these models can generate interpretable rules to guide the adjustment of the cruise controllers’ parameters, they do not improve its performance in low packet reception ratio scenarios. In this regard, an extension to a state-of-the-art Cooperative Adaptive Cruise Controller (CACC) is proposed to account for the possible packet losses of V2X communication technologies. The results show that by using the desired acceleration information not only from the platoon leader and predecessor vehicle, but also from all platooning vehicles, the likelihood of inter-vehicle collisions can be significantly reduced in scenarios with high packet losses. Therefore, it has been demonstrated that, while V2X communication technologies cannot guarantee reliability in high traffic density environments, integrating packet loss considerations into control algorithms can significantly improve the overall reliability of the autonomous driving system.