M. Sc. Jan-Hendrik Ewering

How do can we enable mechatronic systems to be autonomous and intelligent, even though only a limited prior knowledge and few measurement data are available? And how can we fuse the available information in a coherent way? My research focuses on physics-informed machine learning methods to construct a bridge between model-based and learning estimation algorithms. This involves enhancing prior physical knowledge with learning elements in order to obtain interpretable, reliable and data-efficient system models.