Uncertainty Quantification and Propagation in Digital Twins
Cláudio Gomes
Aarhus University
In this talk, I wish to introduce state-of-the-art techniques for uncertainty quantification and propagation and discuss how these can be applied in a digital twin context.
I’ll focus on a small self-adaptation loop (one of the many services a digital twin contains) of an incubator system, and will discuss how uncertainty can be incorporated into the loop, and propagated through its different steps, and inform decision-making (giving meaning to the intuitive statement “if I see highly uncertain traces, then my decisions will be conservative while I wait for more data”.)
The main technique introduced for uncertainty quantification is the Kalman filter.
I’ll conclude with a generalization of this concept, with a discussion on what needs to be provided to enable this technique in other digital twins.
Speaker bio: Cláudio Gomes is an Assistant Professor at Aarhus University, with a record of inter-disciplinary research and with focus on promoting digital twin engineering. His research currently involves uncertainty quantification and propagation in digital twins, as well as application of formal methods within and for digital twins. He is part of the Digit Center, the Functional Mockup Interface standardization committee, and likes chocolate.