graphical interaction representations aiding prognostic health management (GIRaPHMan)
Image source: Department of Defence
Key Points
- DAIRNet Theme: Patterns in Noisy and Dynamic Data
- Organisation: Lockheed Martin Australia
- Point of contact: Dr Leon Clark
The graphical interaction representations aiding prognostic health management (GIRaPHMan) project is investigating the development of an AI framework that allows aircraft health data from multiple sources, such as sensor measurements, service logs, and reporting material, to be combined into a single representation; enabling structural health models to be unified in a common and consistent way.
We propose the use of a semantic graph processing scheme for this purpose. This will naturally accommodate hybrid modelling including the underlying low-level structural dynamics and physics modelling, learned senso-level Prognostic Health Management (PHM) analytics, and higher, token-level sequence modelling.
