Enhancing industrial communication protocols using machine learning

Abstract

This thesis investigates the design, configuration, and performance of industrial communication between a PLC and sensors,actuators.it examines the role of fieldbuses in automation systems. The study compares cyclic and acyclic communication methods, evaluates PROFIBUS data flow and architecture, and explores the integration of machine learning for anomaly detection. Furthermore, the thesis introduces an unsupervised machine learningbased intrusion detection system for PROFIBUS networks, capable of detecting anomalies and potential threats without requiring labeled data. The proposed system demonstrates substantial improvements in real-time industrial control and network security, with particular relevance to factory operations in Algeria