Manufacturing process impacts on occupational health: a machine learning framework

The Operator 4.0 generation denotes a smart and skilled operator accomplishing ‘cooperative work’ with robots, machines and cyber-physical systems. In this taxonomy, a healthy operator is an operator equipped with wearable technology to monitor biometrics in a workplace to monitor and ideally prevent urgent threats to safety, stress in manufacturing and production quality. In a digitalized context, a cloud manufacturing platform for occupational health assessment, capable of collecting physiological, environmental and manufacturing process data can potentially enable prompt action to prevent fatalities. This paper proposes a novel machine learning-based framework and associated methods to classify physiological data acquired using wearable sensors during manufacturing work, to be utilized in a fuzzy-based expert system to determine the level and type of health risk for Operator 4.0. Classification algorithms are presented and a manufacturing case study is illustrated to exemplify the proposed methodology and to evaluate the industrial suitability.