Transfer learning for end-of-line testing and monitoring in fleets
The PhD project will investigate novel methodologies for assessing the performance and usage of assets in a fleet throughout the product lifecycle. A specific application of interest is the end-of-line quality control testing of vehicles, based on NVH (Noise, Vibration, Harshness) measurements. Machine learning and deep learning techniques have the potential to automatically and objectively assess the status of the vehicle based on such measurements. However, these techniques rely on the availability of a sufficiently large training dataset, which may be infeasible to obtain in industrial practice. To overcome the data scarcity problem, transfer learning strategies will be investigated.