Advanced Surveillance Methods in Era of Industry 4.0

In this era of industry 4.0, manufacturing processes are often equipped with highly reliable and advanced manufacturing devices. Thus, manufacturing processes produce a large number of quality characteristics; therefore, data from such processes are known as multivariate data. Often multivariate control charts are deployed to detect changes in the production process as quickly as possible. However, the traditional multivariate control charts are designed under the normality assumption, but this assumption is often violated in most of the production processes. Consequently, traditional multivariate control charts do not implement for real-time monitoring as they lead to high false alarms. Hence, in literature, non-normal multivariate processes are monitored using two approaches: (i) charts based on sign and rank based nonparametric tests and (ii) charts based on exceedance or precedence tests. Many studies are designed based on the first approach in statistical process control literature, but very few studies are available on the later monitoring approach. Usually, the ranked based nonparametric tests perform well when the underlying distribution is closely symmetric. However, the exceedance test outperforms ranked based nonparametric tests when the underlying distribution is skewed. Hence, motivated by this observation, we are intended to propose. This research focusses on the development of multivariate control charts based on the exceedance test for real-time change detection in multivariate dataset. Our proposal can benefit for many applications, for instance, in the health sector to monitor the real-time occurrence of patients health problems. Also, in the manufacturing industry to monitor the breakdown of multiple correlated components/devices of a manufacturing system in real-time. Moreover, our method may also be useful in the service sector, where it will perform the real-time surveillance of employees who are not performing well for increment in industry profit.