Year: 2024
Type: article
Abstract: Cyber-Physical Systems (CPS) enabled by the Internet of Things (IoT) present unique security challenges as security solutions designed for traditional IT/OT systems may not be sufficient in CPS environments. Therefore, in this study, we introduce a two-stage ensemble framework for attack detection and attribution suitable for CPS, specifically industrial control systems (ICS). To identify attacks in imbalanced ICS environments, decision trees are combined with a unique ensemble model for deep representation learning. In a next step, an ensemble of deep neural networks is used to support attack attribution. Datasets from a gas pipeline and a water treatment system are used to practically test the proposed model. The results show that the proposed model performs better than competing methods with the same computational complexity.
Source: Journal of engineering sciences.
Authors: Dr.Ranjith, Mr.Praveen
Institution: Malla Reddy Engineering College for Women
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4405884693
