Physics-informed artificial intelligence for autonomy applications in defence (PERTINENCE)

The PERTINENCE project is dedicated to transferring knowledge and results in the field of physics-informed AI, which combines traditional physics-based modeling methods with AI approaches, from ongoing research in civilian application to defense applications. The methods offer the benefits of AI but within the constraints of the physical system. The project will foster close collaboration between SINTEF and FFI by adopting the methods to defence-specific use cases. The initial focus is on improving online battery status estimation for unmanned aerial systems (UAS) as battery level is a main limiting factor in the operational use of UAS. Current approaches tend to be overly conservative in harsh environments, such as windy and icing conditions. The new methods are expected to enhance energy consumption prediction accuracy, extend UAS operation times, and potentially enable new types of missions. Moreover, these methods have broader applicability in enhancing the performance of defense technologies operating under challenging conditions and uncertainties. As the defense sector increasingly relies on autonomous systems, innovative approaches to address real-world uncertainties are crucial. To ensure uptake of PERTINENCE results, we will disseminate results to the defence sector, as well as the scientific community. Additionally, the project team intends to explore synergies with ongoing FFI projects to further extend the impact of the project, both in terms of knowledge sharing and in terms of establishing new constellations for future collaboration.

The PERTINENCE project is dedicated to transferring knowledge and results in the field of physics-informed AI, which combines traditional physics-based modeling methods with AI approaches, from ongoing research in land-based process industries to defense applications. The methods offer the benefits of AI but within the constraints of the physical system. The project will foster close collaboration between SINTEF and FFI by adopting the methods to defence-specific use cases. The initial focus is on improving online battery status estimation for unmanned aerial systems (UAS) as battery level is a main limiting factor in the operational use of UAS. Current approaches tend to be overly conservative in harsh environments, such as windy and icing conditions. The new methods are expected to enhance energy consumption prediction accuracy, extend UAS operation times, and potentially enable new types of missions. Moreover, these methods have broader applicability in enhancing the performance of defense technologies operating under challenging conditions and uncertainties. As the defense sector increasingly relies on autonomous systems, innovative approaches to address real-world uncertainties are crucial. Dissemination efforts will focus on maximizing impact, targeting FFI and the Norwegian Armed Forces. Based on good experience and positive feedback from similar projects, knowledge transfer to the defence sector will be facilitated by numerous seminars, webinars, and hands-on tutorials. Additionally, the project team intends to explore synergies with ongoing FFI projects to further extend the impact of the project, both in terms of knowledge sharing and in terms of establishing new constellations for future collaboration.