Machine learning-based intrusion detection for SCADA systems in healthcare

Abstract

Energy distribution systems and cyber-physical systems brought together information technology, electrical and mechanical engineering in an integrated manner. This cybernetic-mechatronics development has drawn the attention of both cybercriminals and cybersecurity researchers by expanding the attacks in critical infrastructures. With the development of information communication technology, supervisory control and data acquisition (SCADA) systems will turn into cloud-based systems that can communicate with IoT devices in the future. In addition, SCADA systems can be utilized in hospitals for various aspects and in IoT healthcare environments. However, SCADA protocols communicate on text and do not have a generalized security structure. Intrusion detection systems are structures developed against cyber-attacks that may cause serious damage. These systems try to provide the highest level of security, including both software and hardware structures. In this work, attack detection based on artificial intelligence and machine learning techniques is performed for the classification of attack threats in cyber-physical systems. Intrusion detection based on artificial intelligence and machine learning techniques is performed for the detection and classification of threats against cyber-physical systems. In this context, attack type classification is performed using machine learning algorithms. At the same time, performance evaluation realized by using computational metrics on machine learning algorithms. Attack type determination and performance analysis were carried out in the test environment and the results were discussed.