Industrial IoT Gateway with Machine Learning for Smart Manufacturing

Abstract

Working together is important aspect of future industry. Therefore, technologies like Internet of Things (IoT), cloud computing, SOA give rise to another industrial revolution. We propose here a concept definition, which focuses on data acquisition, integration and predictive control in the industry. The concept consists of industrial IoT gateway, cloud services and machine learning services. We used machine learning to verify our data acquisition solution and we implemented prediction control as a cloud service. Finally, proposed solution will exceed boundaries inside ICS (Information and Control System), improve flexibility, interoperability and test plant prediction control in smart manufacturing.

Edge‐enabled IoT gateway criteria selection and evaluation

“…The edge‐enabled architecture allows smart systems and IoT gateway to provide real‐time system control in some cases. At the SCADA/HMI, control level less than one second from the recording of the event after it was shown is possible …”

“…These criteria fell into four main categories (device connectivity, data pre‐processing, data analytics, and special hardware requirements). For a better understanding, we ranked these criteria as follows: Device connectivity Network standardization and protocols– Services that offer tools to standardize the various protocols coming from connected devices into protocols that are supported and sent to the cloud. This can contain data formatting and message brokers. –Configuration of the connected devices– Tools for organizing connected devices to recognized and user‐defined groups.…”

“…For a better understanding, we ranked these criteria as follows: Device connectivity Network standardization and protocols– Services that offer tools to standardize the various protocols coming from connected devices into protocols that are supported and sent to the cloud. This can contain data formatting and message brokers. –Configuration of the connected devices– Tools for organizing connected devices to recognized and user‐defined groups. This should have a better visibility over the monitored system. Data pre‐processing Data filtering– Tools providing different algorithms that allow different types of filtering to be applied to incoming data without losing the information value of the data. –Data consolidation– Tools providing inbound data aggregation based by topic.…”

“…This can contain data formatting and message brokers. –Configuration of the connected devices– Tools for organizing connected devices to recognized and user‐defined groups. This should have a better visibility over the monitored system. Data pre‐processing Data filtering– Tools providing different algorithms that allow different types of filtering to be applied to incoming data without losing the information value of the data. –Data consolidation– Tools providing inbound data aggregation based by topic. This can contain data fusion and grouping. Data analysis Data pre‐analytics– Tools that provide machine learning algorithms for flowing data analysis and pattern search in these data.…”

“…This should have a better visibility over the monitored system. Data pre‐processing Data filtering– Tools providing different algorithms that allow different types of filtering to be applied to incoming data without losing the information value of the data. –Data consolidation– Tools providing inbound data aggregation based by topic. This can contain data fusion and grouping. Data analysis Data pre‐analytics– Tools that provide machine learning algorithms for flowing data analysis and pattern search in these data. It can be a similar principle to predictive maintenance. –Local database– Local storage of processed data.…”