Securing Cyber-Physical Industrial Systems Against False Data Injection Attacks: A Hybrid Detection Approach

Cyber-physical industrial systems are internet-enabled physical entities embedded with computers and control components consisting of sensors and actuators. However, interconnecting the cyber and physical spaces led to new security challenges. This paper aims to come up with a novel approach to detect False Data Injection Attacks (FDIA) on cyber-physical industrial systems. To reach this goal, we propose an efficient anomaly-based approach for detecting FDIAs against cyber-physical industrial systems. Particularly, we are using both of the physical data and the network traffic data to train two artificial neural networks to detect the FDIAs occurring in the physical and the cyber space. Then, the results of the two ANN are fed to a decision tree to fuse the two detectors in order to improve the detection probability. The extensive simulations prove that our attack detection approach is able to detect attacks with a great accuracy and that this methodology could be very efficient. The proposed approach is compared with existing methods in the literature and evaluated using a hardware-in-the-loop water distribution testbed dataset.