Anomaly Detection on Multivariate Time-Series from Lithography Equipment using Machine Learning

Abstract

The present work is mainly motivated by the challenges embraced by the metallic packing industry, in its path along the fourth industrial revolution (Industry 4.0). This work serves to bring Artificial Intelligence (AI) to a mass production lithography process to detect anomalous patterns, using Outlier Detection (OD) algorithms to prevent non-conformities and support quality control operators. All the OD algorithms deployment is based on Machine Learning (ML) techniques, scratching the surface of the process and quality monitoring applications in industrial scenarios.