The advent of vehicle-to-everything (V2X) communication protocols allows for the development of control strategies achieving fast dissipation of congestion, equitably and autonomously, based on the valuation of travel times of the individuals. Indeed, the ability to consider driver-specific attributes, such as value of time or urgency, in traffic management through advanced communication protocols allows for reducing individual delay costs, leading to economically and performance-efficient outcomes. Under this scope, AUTONOMY proposes an auction-based traffic management paradigm enabled by vehicular communication to more fairly and efficiently allocate roadway space, perceiving drivers as rational, selfinterested economic agents bidding for roadway use. In particular, the project seeks to develop auction-based traffic management schemes for mixed settings of legacy and connected and autonomous vehicle traffic, namely signalized intersection control, uninterrupted flow conditions and autonomous intersection management, thus covering a wide range of potential applications in urban traffic management. The project will leverage second-price auctions, heuristic optimization methods and reinforcement learning to develop new algorithmic frameworks for intersection control and cooperative traffic management. The impacts (including efficiency, equity, sustainability, safety) of the proposed schemes will be subsequently evaluated through different types of applications and scenarios, exploiting open-source traffic simulation platforms for traffic management. Research outcomes will then serve as the basis for scale-up impact assessment and knowledge transfer, through multiple dissemination and stakeholder outreach channels, also resulting in a set of recommendations and readily available tools.
This project is carried out within the framework of the National Recovery and Resilience Plan Greece 2.0, funded by the European Union – NextGenerationEU (Implementation body: HFRI). https://greece20.gov.gr/
How We Contribute
As part of the AUTONOMY project, the DSAIT team is developing innovative AI-driven, market-inspired solutions aiming to improve efficiency, fairness, and safety in future CAV ecosystems across urban and highway environments. Our research focuses on the fairness aspects of auction-based intersection control methods by designing dynamic auctions enhanced by fairness-aware learning agents that ensure equitable and efficient green time allocation. Moreover, the effectiveness of lane changing behavior is revisited regarding platoon formation in highways, and a dynamic regulatory framework is introduced, which combines reinforcement learning and sponsored search auctions towards a flow-aware slot allocation mechanism for vehicle platoons under mixed traffic conditions. Finally, an auction-based bilevel optimization framework is developed for unsignalized intersection control, focusing on platoon-based coordination. More precisely, this framework first determines whether incoming platoons should be split and subsequently calculates the coordination speed for each resulting platoon segment, while accounting for safety constraints and maximizing intersection efficiency. All solutions are being integrated into simulation environments for performance validation under various traffic scenarios, supporting AUTONOMY’s vision of future intelligent traffic management.