Design and Implementation of an Embedded System for Real-Time Condition-Based Monitoring and Anomaly Detection

Sammanfattning

This project presents the design and implementation of an embedded system for real-time data acquisition and handling from sensors in an industrial environment. The system was designed to collect, process, and visualise data from multiple sensors using IO-Link communication. The system consists of a Raspberry Pi and an IO-Link master, with the MQTT protocol used for communication. Sensor data was sent from the IO-Link master to the Raspberry Pi for processing. This was done using statistical methods such as rolling averages and standard deviations for anomaly detection. The data was then presented to the user using a graphical user interface. This enabled real-time monitoring and alert generation. The results show that the constructed system successfully handled continuous data streams of two data points per second from the three sensors. The system also detected deviations introduced during testing by detecting values that deviated from the norm and irregular behaviour. The artefact operates independently of the existing hardware and does not have a cloud service connection. This project demonstrates an approach to IoT-based condition monitoring and presents the potential for future development towards scalable data storage and predictive maintenance for the system.