- Learning-Based Optimal Control with Performance Guarantees for Unknown Systems with Latent States. 2024 European Control Conference (ECC), 2024, 90-97 mehr… BibTeX Volltext ( DOI ) Volltext (mediaTUM)
- Robust Safe Learning and Control in an Unknown Environment: An Uncertainty-Separated Control Barrier Function Approach. IEEE Robotics and Automation Letters 8 (10), 2023, 6539-6546 mehr… BibTeX Volltext ( DOI )
- Vision-Based Uncertainty-Aware Motion Planning Based on Probabilistic Semantic Segmentation. IEEE Robotics and Automation Letters 8 (11), 2023, 7825-7832 mehr… BibTeX Volltext ( DOI ) Volltext (mediaTUM)
- Safe Planning and Control Under Uncertainty: A Model-Free Design With One-Step Backward Data. IEEE Transactions on Industrial Electronics 71 (1), 2023, 729-738 mehr… BibTeX Volltext ( DOI ) Volltext (mediaTUM)
- Deep Learning based Uncertainty Decomposition for Real-time Control. 2023, The 22nd World Congress of the International Federation of Automatic Control, 2023 mehr… BibTeX Volltext (mediaTUM)
- Safe Learning-Based Control of Elastic Joint Robots via Control Barrier Functions. 2023, The 22nd World Congress of the International Federation of Automatic Control, 2023 mehr… BibTeX Volltext (mediaTUM)
ONE MUNICH Strategy Forum Project - Next generation Human-Centered Robotics
In the ONE MUNICH Strategy Forum Project “Next generation Human-Centered Robotics”, 17 research groups from the three leading Munich research institutes, Technical University of Munich, Ludwig Maximilian University of Munich, and Helmholtz Zentrum München, conduct joint research on various aspects of machine intelligence, systems theory, and translational medicine. The project aims to leverage their collective expertise and efforts to develop novel technologies to address major societal health challenges.
Motivation
An aging society and increasingly complex care present significant challenges to the healthcare system. Novel technologies, particularly in cooperative robotics, offer the possibility of overcoming these problems. For example, patients could be treated more quickly and efficiently via teleoperation utilizing robotic avatars and wearable haptic devices. Similarly, amputee patients could resume everyday tasks themselves with "seamless" prostheses controlled by brain signals and providing sensory feedback to the user. However, to make this vision a reality, many obstacles in the areas of trustworthy and embodied AI and human-robot interaction still need to be overcome. The ONE MUNICH project aims to bring together different research groups with highly complementary expertise to holistically address these problems and develop creative solutions in interdisciplinary teams.
Research focus at ITR
ITR contributes its experience in safe and robust learning-based modeling and control to the project. In particular, we are involved in a research field that deals with mathematical frameworks for trustworthiness in learning-facilitated estimation and control. While learning-based algorithms for perception and control of autonomous systems have achieved tremendous success in recent years, few approaches can provide the rigorous robustness guarantees required for safety-critical applications. Therefore, this research area aims to develop new trustworthy learning-based algorithms and mathematical tools for their analysis.
ITR team members
- Sandra Hirche (Principal Investigator)
- Robert Lefringhausen
Funding program
This research project is supported by LMUexcellent and TUM AGENDA 2030, funded by the Federal Ministry of Education and Research (BMBF) and the Free State of Bavaria under the Excellence Strategy of the Federal Government and the Länder as well as by the Hightech Agenda Bavaria.