Attacks Detection in Industrial Cyber-Physical Systems Using Convolutional Neural Networks

Cyber-physical systems (CPSs) are widely used and extremely important due to their promise for substantial and long-term benefits to society, economy, environment, and human life. Moreover, the development in communication, computing, and storage technologies has resulted in a revolution in information communication technologies (ICT). The utilization of CPSs in industrial control systems are well known as industrial CPSs (ICPSs). Consequently, these systems have become a popular target for cyber-attacks and malicious threats which can disable the system’s functioning and have serious safety-related consequences. This paper presents an attack detection method based on simple neural networks, 1D convolutional neural networks. The presented method is verified using a popular public dataset, the Secure Water Treatment testbed (SWaT), which is a small-scale representation of a real-world industrial water treatment plant. The results have demonstrated the effectiveness of the presented method for attack detection in ICPSs.