We are excited about our two contributions to the 2024 63rd IEEE Conference on Decision and Control (CDC) in Milan. Congratulations to the authors!
Efficient Online Inference and Learning in Partially Known Nonlinear State Space Models by Choosing Expressive Degrees of Freedom. Jan-Hendrik Ewering, Björn Volkmann, Simon F. G. Ehlers, Thomas Seel, and Michael Meindl
In this paper, we present an algorithm to enable efficient estimation and learning in real-world systems. The unknown model components are learned during operation solely based on input and output data. The key element that makes this efficient and achievable without expert knowledge is the preconditioning of the learning-based model components, ensuring that the learning during operation is restricted to the most expressive degrees of freedom.
Reference-Adapting Iterative Learning Control for Motion Optimization in Constrained Environments. Michael Meindl, Simon Bachhuber, Thomas Seel
In this paper, we present a methodological extension of conventional Iterative Learning Control, enabling robots to autonomously optimize their motion execution even in constrained environments.
We would like to thank the entire team, especially Jan-Hendrik Ewering and Michael Meindl, for their dedicated work on these projects. Your collaboration and hard work have played a crucial role in this success.
You can find more information here: https://cdc2024.ieeecss.org/