At the BMW Summer School 2025 in southern France, Ramakrishnan Subramanian presented his research on the safety-critical limits of autonomous vehicles.
Around 30 selected doctoral students from around the world gathered in Saint-Raphaël from 29 June to 4 July 2025 for the eleventh BMW Summer School, under the theme of ‘Trust and Security in Artificial Intelligence’. Among them was Ramakrishnan Subramanian, a research associate in the inIT working group 'Intelligent Systems', which is headed by Prof. Dr. Ulrich Büker.
Exchange between science and industry
The BMW Summer School is organised annually by BMW in cooperation with the Technical University of Munich (TUM) and EURECOM. It offers doctoral students whose research focuses on intelligent mobility a platform for engaging in dialogue with leading industry and scientific minds. This year's key topics included AI and functional safety, generative AI services, chip design for automated driving, and trust and safety in AI.
Ramakrishnan Subramanian presented a poster entitled ‘When Autonomous Vehicles Hit Their Limits', which explored the limitations of autonomous driving systems in real-world scenarios. The focus was on how automated vehicles react to unsafe conditions that arise due to situations they are not designed for, out of which few important use cases include slippery roads, poor visibility affecting the sensors or temporary events on roads like construction zones or closures due to accidents – and how such situations can be detected early and safely controlled.
Practical learning in a start-up format
A particular highlight of the week was the Lean Startup Machine, in which participants worked in interdisciplinary teams to develop a viable business model based on their research topics, with support from BMW innovation managers. Ramakrishnan Subramanian's team won the Best Pitch Award with their 'Eyeway' concept.
‘The Summer School was a unique opportunity to discuss my topic with international experts,’ said Ramakrishnan Subramanian. ‘Above all, the discussion of industrial requirements for autonomous systems and the close connection to practical application provided valuable impetus for further developing my doctoral thesis.’