Embodied AI · Computer Vision · Machine Learning
My work centers on scalable training data generation and real-world deployment of end-to-end driving stacks — bridging HD maps, LiDAR, and camera perception to train robust real-world models under limited annotation. I've led research groups, served as principal investigator on six-figure industrial grants, published across NeurIPS, CVPR, IROS, RAL, T-ITS, and T-IV, and deployed perception systems on autonomous research vehicles across multiple European cities.
I lead the Visual & Spatial Learning group at the FZI Research Center for Information Technology, where my team works on auto-labeling pipelines for multimodal perception and end-to-end driving systems. I completed my PhD at the Karlsruhe Institute of Technology (KIT), supervised by Prof. Christoph Stiller, defending my thesis "Learning from Maps: Scalable Ground Truth Generation in Autonomous Driving" in March 2026 with Summa cum Laude (1.0/1.0) distinction.
My research bridges maps, perception, and learning to enable robust real-world AI models under limited annotation and significant sensor domain shifts. I am the E2E-owner and co-adviser for the KITScenes Multimodal dataset, and I am particularly interested in leveraging HD maps and multi-drive mapping to overcome data scarcity in autonomous driving.
I co-maintain the software and hardware stacks for up to two autonomous driving research vehicles, and have developed strong proficiency in Python, C++, ROS, and modern deep learning frameworks (PyTorch, TensorFlow). I currently supervise 7 PhD students after having supervised 11 MSc thesis projects.
I actively collaborate with leading research institutions and industry partners worldwide, bridging fundamental research with real-world industrial autonomous driving applications.
My research focuses on enabling robust and scalable embodied AI systems for real-world applications, with emphasis on autonomous driving and mobile robotics.
Scalable ground truth generation for autonomous driving by leveraging high-definition maps and multi-drive data to overcome annotation bottlenecks and enable data-efficient learning.
Developing robust domain adaptation techniques for embodied AI systems to generalize across different sensor modalities, environmental conditions, and deployment scenarios.
Scaling real-world deployment of end-to-end autonomous driving stacks in complex urban environments, from perception through planning to control.
Frank Bieder, Hendrik Königshof, Haohao Hu, Fabian Immel, Yinzhe Shen, Jan-Hendrik Pauls, Christoph Stiller
Cross-modal domain adaptation for robust perception across different sensor modalities.
Richard Schwarzkopf et al. (incl. Frank Bieder)
Large-scale multimodal dataset for autonomous driving research. Frank is E2E-owner and co-adviser.
Fabian Immel, Jan-Hendrik Pauls, Richard Fehler, Frank Bieder, Jonas Merkert, Christoph Stiller
Online HD map construction using text-annotated navigation maps.
Fabian Immel, Richard Fehler, Frank Bieder, Jan-Hendrik Pauls, Christoph Stiller
Generalist transformer model for real-world high-definition map completion.
Frank Bieder, Haohao Hu, Johannes Schantz, Oguzahn Kirik, Florian Ries, Martin Haueis, Christoph Stiller
Automated training data generation using HD maps and multi-drive mapping.
Sven Richter, Frank Bieder, Sascha Wirges, Christoph Stiller
Evidential top-view representation for modeling the semantic environment around automated vehicles.
Haohao Hu, F. Han, Frank Bieder, Jan-Hendrik Pauls, Christoph Stiller
Targetless extrinsic calibration of multi-modal sensors with uncertainty quantification.
Christoph Burger, Johannes Fischer, Frank Bieder, Ömer Tas, Christoph Stiller
Game-theoretic approach to motion planning considering interaction between automated vehicles.
Kunyu Peng, Juncong Fei, Kailun Yang, Alina Roitberg, Jiaming Zhang, Frank Bieder, Philipp Heidenreich, Christoph Stiller, Rainer Stiefelhagen
Multi-attentional approach for dense semantic understanding from LiDAR data.
Frank Bieder, Maximilian Link, Simon Romanski, Haohao Hu, Christoph Stiller
Novel approach to semantic segmentation using complementary feature representations.
Frank Bieder, Sascha Wirges, Johannes Janosovits, Sven Richter, Zheyuan Wang, Christoph Stiller
Multi-layer grid map representation for robust semantic segmentation from sparse sensor data.
FZI Research Center for Information Technology
Visual and Spatial Learning for Autonomous Driving · Supervised by Prof. Christoph Stiller
FZI Research Center for Information Technology
Visual and Spatial Learning for Autonomous Driving · Supervised by Prof. Christoph Stiller
UC Berkeley – Mechanical Systems Control Lab
Supervised by Prof. Masayoshi Tomizuka
Atlatec GmbH (acquired by Bosch in 2022)
Supervised by Dr. Julius Ziegler
Daimler AG
Supervised by Dr. Uwe Franke & Fiete Botschen
FZI Research Center for Information Technology, Karlsruhe
Supervised by Sascha Wirges
University of Ottawa VIVA Lab, Ottawa, Canada
Supervised by Robert Laganière
Mercedes-Benz Research & Development, Sunnyvale, USA
Supervised by Jörg Hillenbrand · MBRDNA's Vehicle Intelligence Group
Karlsruhe Institute of Technology (KIT)
Thesis: Learning from Maps: Scalable Ground Truth Generation in Autonomous Driving
Supervisor: Prof. Christoph Stiller · Defended March 2026 · Summa cum Laude (1.0/1.0) with Distinction
UC Berkeley – Mechanical Systems Control Lab
Research on map-less driving and map perception · Supervised by Prof. Masayoshi Tomizuka
Funded by KHYS Exchange Grant
Karlsruhe Institute of Technology (KIT)
GPA: 1.4/1.0 · Focus: Machine learning, system theory & robotics
Thesis: Supervised Domain Adaptation for Semantic Segmentation (Grade: 1.0/1.0)
University of Ottawa, Canada
Canadian GPA: 3.8/4.0 · Focus: Machine learning and data processing
Karlsruhe Institute of Technology (KIT)
Thesis: Fusion of Data from Stereo Camera and Radar Sensor for Tracking of Extended Objects (Grade: 1.0/1.0)
German (native), English (fluent), French (upper intermediate)
Python (PyTorch, TensorFlow, OpenCV, NumPy), C++, ROS, containerized workflows
Supervised 11 MSc thesis projects · Currently co-supervising 7 PhD students
1st Organizer: FMAD – Foundation Models for AD (ITSC 24/25), MB2ML – Mapless AD (IV 22/23)
Completed MBA Fundamentals Program, Hector School of Engineering and Management
Engineers Without Borders (2012–2013) · International Student Mentoring (2019–2022)
I'm always happy to discuss research collaborations, student projects, and opportunities in autonomous driving and machine learning. Feel free to reach out via email or connect on social media.
Open to: Research collaborations · PhD student supervision · Master's thesis supervision · Industrial partnerships · Speaking engagements
