Real-time Data Analytics Edge Computing Application for Industry 4.0: The Mahalanobis-Taguchi Approach

Abstract

Industry 4.0 and its innovative technologies (e.g., Internet of Things, Cyber-Physical Systems, Cloud Computing, Big Data and Artificial Intelligence) represent great promise. Still, com-panies experience hardship when transforming from reactive to predictive manufacturing systems. The latter, driven by data science development, use predictive models to detect and solve production and maintenance issues before they happen. To eliminate the need for large and varied datasets for development of predictive models, in the present research we propose development of real-time predictive models based on small dataset without faulty data. This is achieved by using Mahalanobis-Taguchi system for fault detection in lack of fault data samples, and by using Edge Computing environment which provides higher re-sponsiveness, better security and decreased costs. Subsequently, two predictive models are developed, tested and compared for the case company from process industry (i.e. the vi-nyl-floor industry sector). Finally, recommendations for the industry are provided.