PhD in Machine Learning

The aim of this PhD project is to develop explainable physics-informed RNNs for autonomous navigation and neural observer design within the broader framework of Embodied AI. The goal is to integrate physical models with deep learning to create interpretable, data-driven observers that enable physically grounded perception and control for robust state estimation and bias correction under uncertain sensing conditions.

The PhD candidate will be supervised by Professor Thor I. Fossen, with Assoc. Professor Erlend M. Lervik Coates as co-supervisor.