Architecture of IoT Sensor Data Acquisition Systems and IoT Data Processing for Anomaly Detection

Today's industry is increasingly characterized by the integration of Internet of Things (IoT) devices and the rapidly spreading digitization trend, which are also known as the foundations of Industry 4.0. The implementation of production automation also become an important issue, as well as the continuous collection of data and their storage in the cloud. As a result of these integration, the method of collecting data, the usability of cloud-based systems, the feasibility of artificial intelligence-based data analysis, the visualization of massive amounts of collected data, cyber security issues, etc. have become everyday issues. Industry players must devote significant attention and resources to real-time data processing to extract vital information from available datasets. This includes identifying outlier data, filtering fake information, and enabling predictive maintenance through forecasting analysis. In this article, we provide a detailed overview and the general architecture of IoT (Industrial IoT) systems, its various cloud service-based implementations, and present our alternative solution built using open-source software components. At the same time, we present some showcases and prototype applications using our system for outlier detection and predictive model building based on the collected sensor data.