Mission
Our research focuses on machines that coexist with humans in physical, social, and emotional spaces, with a particular emphasis on AI-enabled interactive robot systems. Our work lies at the intersection of robotics, machine learning, and human-robot interaction.
Research
Our research advances Open-World Embodied Intelligence, where we use, e.g., multimodal foundation models to enable open-vocabulary zero-shot scene understanding for tasks such as 3D scene graph learning, affordance grounding or scene question answering. For Engaging Interactive Robots, we develop methods that turn non-social robots into socially intelligent machines that maintain long-term human engagement, e.g., through expressive non-verbal behavior and generative AI-driven design. Towards AI-Based Industrial Robots, we integrate model-based and data-driven techniques to achieve flexible, autonomous performance in complex real-world tasks like bin packing, in close collaboration with industry partners such as the Fraunhofer Society.
Teaching
We teach all aspects of robotics and intelligent systems, from the classical foundations of the field, to state-of-the-art theories, methods and experiments, to practical courses with hands-on programming and integration experience. We offer specialized modules for both Bachelor’s and Master’s students
Partners
Prof. Arras is a Cyber Valley Endowed Chair for Autonomous Systems, Europe’s largest and leading center for excellence in AI and modern robotics.
The BMBF-funded Robotics Institute of Germany is a unique initiative that connects the 14 German universities and research institutions that conduct internationally leading research in the field of AI-powered robotics.