During my studies of Engineering Cybernetics I wondered why it is so hard to reach high control performance on real systems. It seemed so easy in the simulations! The problem is that the real world is never perfect: Imprecisions in the mechanics, noise in the sensors and model errors make it hard to build good controllers.
Therefore, my focus lies on the probabilistic view on controllers as well as the connections between control theory and machine learning.
Uncertainties have long been recognized as a key difficulty for control, deteriorating performance or even putting system safety at risk. This issue has been classically addressed by robust controller design, making use of a deterministic bound on the uncertainty and designing the controller for all possible uncertainty realizations...
Our goal is to understand the principles of Perception, Action and Learning in autonomous systems that successfully interact with complex environments and to use this understanding to design future systems