Optimization for failure detection in autonomous driving

This project is together with NXP on a Digital Twin for lifetime prognostics to cover failure modes related to packaging (e.g. bondwire cracks or solder joint fails). Within the concept of the digital twin a connection between the virtual model (e.g. finite element model, response surface model, or analytical model) and the physical product is established.

The aim is to develop an, experimentally validated, digital twin to monitor package related fails. Based on large sets of ageing data, a dataset containing the relevant product specifications will be created and processed using a statistical approach. Data-driven methods combining state-of-the-art data analytical methods from AI, and statistical degradation models will be used. Finally, validated digital twin driven prognostics of lifetime for the use case, application of data analysis and modelling techniques to application level digital twin can provide a virtual space of the system to predict the behavior of real entity.

For more information: click here

Requirements

Creative and motivated MSc students Background in microelectronics, material science or nanotechnology.

This project is connected to the EU project Architect, funding by NXP is possible.

Contact

prof.dr.ir. Willem van Driel

Electronic Components, Technology and Materials Group

Department of Microelectronics

Last modified: 2020-01-22