Generating Attacks and Labelling Attack Datasets for Industrial Control Intrusion Detection Systems
Nicholas R. Rodofile · Queensland University of Technology · 2018
To address the arising Cyber Security threats against SCADA-based Critical infrastructure, the security research community have identified the application of Intrusion Detection and IA as an ideal security measure for such systems. The research presents a cyber-attack classification for critical infrastructure, to identify the cyber-attack landscape for critical infrastructure attacks. To further aid in the development and evaluation of AI using intrusion detection, the thesis presents a SCADA cyber-attack generation framework. The cyber-attack generation framework provides a collection of algorithms to stimulate control system equipment with cyber-attacks. Using the attack generation methodology, a SCADA attack labelling framework is also presented to generate labelled attack datasets. The datasets can be used in future work to aid in the development of AI detecting new and unknown cyber attacks on Critical Infrastructure systems.