news

Members of our Lab are keeping busy during the conference season. Today, Onur is presenting his work titled “Routing Autonomous Vehicles Using RL – Progress and Future Directions” at the 18th European Workshop on Reinforcement Learning (EWRL 2025) in Tübingen. This is another edition with our participation, and therefore another opportunity for broad discussion on open problems and future research directions within the European reinforcement learning research community.

Last week Hanna participated in the SUM project General Assembly in Fredrikstad. The two days were marked by discussions on the progress of the project and provided a platform for the mutual exchange of experiences. We also reviewed the achievements to date within Living Labs and among our partners. Additionally, we set the priorities for the third year of the project – all with the aim of delivering tangible benefits to cities across Europe 🌆💪 Many thanks to the organizers for the excellent organization!

Today, during the 10th International Symposium on Dynamic Traffic Assignment (DTA2025) in Salerno Grzegorz, on behalf of Gunnar Flötteröd, Rafał Kucharski and David Watling, is presenting the work Coexistence of human drivers and CAVs - behavioral models and long-term properties. DTA is dedicated to the latest research and methods for modeling dynamic traffic assignment, which help better understand and predict road users’ behavior in real time. This year’s edition offers a valuable opportunity for Grzegorz to share insights on cutting-edge optimization, simulation, and analysis methods for transportation systems aimed at enhancing efficient traffic management.

During Euro Working Group on Transportation 2025 27th Annual Conference Olha is presenting our research on how to strategically select areas for feeder service to maximise a shift from private cars to public transport. These results are a part of SUM Project. Such a great opportunity to get feedback and exchange experiences.

Last two weeks of July were very busy time for Anastasia who attended the DLRL Summer School 2025 in Edmonton, Canada. It allowed to expand her knowledge on reinforcement learning. A highlight was the keynote by Richard Sutton. She also worked with Lucas Cruz on a project based on an idea she proposed, which they presented on the final day. We are glad we could support her visit!

We’re thrilled to announce that the team’s latest publication has been accepted for publication in the highly ranked journal SoftwareX. Congratulations to the entire team — a truly well-deserved success! 👏 You’ll soon be able to read about RouteRL in open access. RouteRL is a new framework that combines multi-agent reinforcement learning with detailed traffic simulations to explore route choice strategies for autonomous vehicles. It models both human drivers and AVs, enabling research in transportation, MARL, and human-AI interaction. This study introduces the framework, its research potential, and illustrative use cases.