Learning controllers for machines: Paradigms and recent results (Prof. Sebastian Trimpe, Institut für Data Science in Maschinenbau, RWTH Aachen University)
Systems & Control Seminar
Abstract
Contacts, nonlinearities, and tedious tuning are just a few of the many reasons we are interested in using learning in the design of controllers—all of which are commonly encountered in robotics and other physical machines. In this talk, we will approach the problem of learning controllers through the lens of three paradigms: deep reinforcement learning, automatic imitation learning from optimal control, and auto-tuning via Bayesian optimization. I will highlight some of our recent results motivated by key challenges faced in practice, such as leveraging uncertainty for better reliability in model-based reinforcement learning and parameter-adaptive approximate model predictive control for imitation learning without retraining. Finally, by discussing these recent advancements alongside applications—showcased through videos of learning robots and machines—I aim to convey an understanding of the potential of these paradigms in both research and current practice.
Biographical information
Sebastian Trimpe is a Full Professor at RWTH Aachen University, where he heads the Institute for Data Science in Mechanical Engineering (DSME) since May 2020. Additionally, he is a founding director of the RWTH Center for Artificial Intelligence and, as of 2023, serves as one of its two Executive Directors. Sebastian's research focuses on fundamental questions at the crossroads of machine learning, control, networked systems, and robotics, with innovative applications thereof. Prior to his role at RWTH, he was a Max Planck Research Group Leader at the Max Planck Institute for Intelligent Systems in Tübingen/Stuttgart. Sebastian earned his Ph.D. degree in 2013 from ETH Zurich, working with Raffaello D'Andrea at the Institute for Dynamic Systems and Control. His academic journey includes a B.Sc. in General Engineering Science (2005), an M.Sc. (Dipl.-Ing.) in Electrical Engineering (2007), and an MBA degree in Technology Management (2007), all from Hamburg University of Technology. Sebastian is recipient of the triennial IFAC World Congress Interactive Paper Prize (2011), the Klaus Tschira Award for achievements in public understanding of science (2014), the Best Paper Award of the International Conference on Cyber-Physical Systems (2019), and the Future Prize by the Ewald Marquardt Stiftung for innovations in control engineering (2020).
Termin
28. Jan. 202516:00 - 17:00
