Anomaly Detection in Industrial Control Systems using Machine Learning
Pradeep Kumar, S Senthil Pandi, Lokesh Kumar, R Karthick
Abstract
Industrial Control Systems (ICS) are vital components utilized in a number of sectors, such as manufacturing, transportation, and energy. Anomaly detection is crucial to guaranteeing the security and dependability of these systems since they are becoming the target of cyberattacks and operational malfunctions. In order to detect and address possible risks and operational anomalies, this study suggests a machine learning-based method for anomaly detection in ICS. Our method combines supervised and unsupervised machine learning methods to identify anomalies in real-time by utilizing past ICS data. To improve detection accuracy and scalability, the suggested system architecture combines modules for anomaly detection, feature selection, data preprocessing, and display. The system’s effectiveness is validated by experimental results, which show that it achieves good precision and recall across a variety of datasets. The suggested system helps create safer and more robust industrial settings.