Developing an efficient attack detection model for an industrial control system using CNN-based approaches : attack detection using PS-CNN
Mohammed, Al-Humairi (2023)
Mohammed, Al-Humairi
2023
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2023052514016
https://urn.fi/URN:NBN:fi:amk-2023052514016
Tiivistelmä
Industrial Control Systems (ICS) critical infrastructures, including thermal plants, water treatment plants, nuclear plants, oil refineries, and gas pipelines, rely on uninterrupted operations. With the transformation of ICS from proprietary to open architectures, these systems are exposed to novel threats and cyber-attacks. Detecting cyber-attacks in ICS at an early stage is crucial to protect critical infrastructures (CIs) from major attacks. Researchers have focused on adapting Information Technology (IT) solutions to enhance ICS security, but ICS-specific attack vectors are often overlooked. This thesis addresses the need for an intelligent attack detection model in ICS, specifically designing the PS-CNN (PCA Select KBest CNN) model. The model demonstrates high accuracy in attack detection and reduced false positives by utilizing feature extraction and selection techniques. The proposed model, evaluated using the UNSW-NB15 dataset, employs Convolutional Neural Networks (CNNs) for classifying malicious attacks in ICS. Comparative analysis validates the effectiveness of the proposed model, highlighting its superiority over existing techniques.