We are incredibly proud to share that our former student Andreas Blum has been awarded the WITTENSTEIN-Price for the best Master's thesis at the TUM School of Engineering & Design!

Andreas worked on the topic of end-to-end differentiable SLAM for autonomous vehicles together with PhD student Felix Jahncke and Prof. Johannes Betz. Throughout the project, he showed incredible thoughtfulness, initiative, and dedication.

The topic matters. Classical, rule-based SLAM is robust and easy to interpret. Fully end-to-end, neural network approaches are flexible and powerful. We found a way to combine the strengths of both into a single, fully differentiable system. Using end-to-end learning, it adapts to new sensors and new conditions. And it runs in real time on compact, energy-efficient hardware.

We are incredibly proud of Andreas and the work our lab is doing in this area, and there is a lot more to come.