Intrusion detection system for IIoT and IoT networks based on machine learning approaches

Abstract

newline It is proposed in this thesis the design of an IDS for an ICS system to detect attacks and raise an alarm to prevent attacks. Our interest stems from the recent increase in cyber security events impacting devices related to IoT. This study aims at assessing ML techniques which have been utilized all through task of identifying intrusions in ICS systems. For this purpose, it was used a dataset generated by an ICS testbed, which consists of an IIoT system to monitor the level of water and turbulence of the storage reservoir. Advanced assaults were launched against an ICS testbed. ML modes were employed to identify assaults in the SCADA system, and their performance is evaluated using various metrics. The results demonstrate excellent execution of machine learning techniques in recognizing intrusions, and also a low number of false notification rate, inferring that it classifies regular data packets quite effectively.