Ralf Römer

prof_pic.jpg

I am a PhD student at the Technical University of Munich (TUM) in the Learning Systems and Robotics Lab (LSY), advised by Prof. Angela Schoellig. Recently, I spent time as a visiting researcher at ETH Zurich in the Learning and Adaptive Systems Group, advised by Prof. Andreas Krause. My research focuses on physical AI: enabling robots to safely and autonomously improve throughout their operational lifetimes, making them reliable and effective for deployment in real-world settings. To this end, I develop methods for uncertainty-aware post-training and continual learning for generalist robot policies (e.g., VLAs, WAMs).

Before joining LSY, I obtained a Master’s degree in Electrical and Computer Engineering at TUM in 2023 and a Bachelor’s degree in Mechatronics at FAU Erlangen-Nuremberg in 2020, funded by the German Academic Scholarship Foundation. During my studies, I spent a semester at EPFL and conducted research in machine learning for robotic throwing with Prof. Aude Billard. Previously, I also worked with Prof. Sandra Hirche at TUM and with Prof. Knut Graichen at FAU. In 2021, I did a research internship in optimal control for autonomous driving at Bosch Research in Renningen, Germany.

Student Supervision: I am always open to supervising excellent and ambitious Master and Bachelor students for a semester project or thesis. You can find a non-exhaustive list of potential topics here. If you want to work with me, please send me an email describing your area of interest and attach your CV and up-to-date transcripts.

News

Jul 2026 Gave an invited talk at the RSS Workshop on Trustworthy Embodied Foundation Models. At the Diffusion for Robot Learning Workshop, we received the best paper award! Impressions.
Jun 2026 Gave an invited talk at the Robot Learning Lab at the University of Freiburg. Impressions.
Apr 2026 Our paper “CLARE: Continual Learning for Vision-Language-Action Models via Autonomous Adapter Routing and Expansion” has been accepted to RA-L. Website.
Apr 2026 Started as a visiting researcher at the Learning and Adaptive Systems Group at ETH Zurich, hosted by Prof. Andreas Krause.
Mar 2026 Our paper “CRISP - Compliant ROS2 Controllers for Learning-Based Manipulation Policies and Teleoperation” has been accepted to RA-P. Website.

all news →

Selected Publications

  1. Preprint
    2026_UQ-VLA.png
    Uncertainty Quantification for Flow-Based Vision-Language-Action Models
    Ralf Römer, Maximilian Seeliger, Saida Liu, Ben Sturgis, Marco Bagatella, Daniel Marta, Andreas Krause, and Angela P. Schoellig
    arXiv preprint, 2026
  2. RA-L
    2026_CLARE.png
    CLARE: Continual Learning for Vision-Language-Action Models via Autonomous Adapter Routing and Expansion
    Ralf Römer*, Yi Zhang*, Yuming Li, and Angela P. Schoellig
    IEEE Robotics and Automation Letters (RA-L), 2026
  3. NeurIPS
    2025_FIPER.png
    Failure Prediction at Runtime for Generative Robot Policies
    Ralf Römer*, Adrian Kobras*, Luca Worbis, and Angela P. Schoellig
    In Advances in Neural Information Processing Systems (NeurIPS), 2025
  4. RA-L
    2025_Semantic-Safety.jpg
    Semantically Safe Robot Manipulation: From Semantic Scene Understanding to Motion Safeguards
    Lukas Brunke, Yanni Zhang, Ralf Römer, Jack Naimer, Nikola Staykov, Siqi Zhou, and Angela P. Schoellig
    IEEE Robotics and Automation Letters (RA-L), 2025