A Neural Network Model for Quality Prediction in the Automotive Industry
Résumé
In this paper, we present a machine learning application for the automotive industry. We study the use of neural networks to predict the location of milled holes in a bumper beam using historical measurement data. The overall goal of the study is to reduce the time needed for quality control procedures as the predictions can supplement manual control measurements. Our preliminary results indicate that the neural network can generally capture the production process variations, but underestimates larger deviations from the specified location.
Domaines
Informatique [cs]
Origine : Fichiers produits par l'(les) auteur(s)