Luentosarja: Data-Model Fusion For Digital Twins on Crewed and Crewless Vessels

This event is part of a lecture series for experts, professors and industry leaders in the maritime sector

The ability to use numerical engineering models to support ship operations, as well as design, has resulted in the concept of the digital twin.  In such approaches, measurements from real-world responses are used to perform prognosis for future events.  However, idealized engineering models seldom agree with real-world observations, leading to uncertainty in the resulting prediction.  This lecture will briefly review types of digital twins, and then present research on using Bayesian networks, neural networks, and Gaussian process machine learning approaches to improve the performance of digital twin prognosis systems.

Lecturer: Matthew Collette

Read more