A Machine Learning and Deep Neural Network Approach in Industrial Control Systems

Abstract

Abstract The two major components in monitoring and controlling industrial processes are SCADA (Supervisory Control and Data Acquisition) and ICS (Industrial Control System). Since their demand has been increased all over the World, they have gained more attention and because of their efficiency and high performance, these systems became mandatory in all countries. In recent times it has become an interesting target for adversaries because most of the infrastructure is automated but security is not strong enough to protect the entire system. Many loopholes make the attacker a way to exploit the vulnerability, hence protecting these systems is becoming more critical. A Hardware-in-the-loop testbed is used in this paper and its main purpose is to simulate power generation units and various attacks are performed on this testbed as well as the attack dataset is also exploited. In this research paper different machine learning algorithms are applied to the dataset and it is found that AdaBoost has better accuracy and performance compared to other algorithms and when it comes to deep learning CNN has the best accuracy compared to other ones.KeywordsIndustrial control systemsSupervisory control and data acquisitionDeep learningMachine learningData mining