Starting a PhD

Last month, I started a PhD. I was interviewing for a research and development engineering position, but only found development and management ones. In France, a PhD is a three years degree, but it is mostly constrained by funding.

For the past month, I tried to fix the reasearch field with bibliographical research. Objectives are to close loop control any injection molding machine to optimize the product quality.

My thesis will be part of the SAPRISTI FUI project, funded by BPI France. Two international automotive injection molders are in, with two research labs and two smaller compagnies.

Development of a methodology for dimensional control of injected parts

Directors: Eric Pairel, Maurice Pillet [theses.fr]
Grenobles Alpes University, SYMME laboratory

The SAPRISTI project aims to develop a self-correcting injection molding system for the production of plastic parts based on product inspection. This system will work autonomously in order to stabilize the process quality correcting aberrations.
This project seeks the optimal control of the quality of parts produced in injection molding.
My research is about inline quality analysis, final products quality prediction and optimal machine control.

The main limit seems to be automated quality control. Maybe Deep Learning could helps. I will try to integrate volumetric 3D scanning and multi-spectral imagery in the short industrial production cycle time.

About the closed loop control, maybe I could do better than Model Predictive Control, with recent deep Reinforcement Learning methods.

Finally, the main problem will be to design a quality control system, with solution for security and ethical concerns.