Development of a Secure Networked Control Algorithm for Industrial IoT Systems Against Cyber-Attacks

Abstract

There are a number of IIoT devices being integrated with existing NCS. This has greatly increased the attack surface of industrial automation infrastructure. Here we propose a secure network control algorithm that allows the operation and stability of closed-loop control systems when subjected to both FDI and DoS attacks. The new algorithm performs anomaly detection based on an observer model and compensates for attack-induced corruptions in both sensor and actuator data. Evaluation of the algorithm was completed through use of a simulated testbed representing an IIoT enabled networked process-control loop that is subject to various communication layer cyber-attacks. The simulation results showed that the proposed algorithm achieved a detection rate of 97.3%, delivered a false alarm rate of just 2.1%, and limited the mean absolute error of the control to below 1% under maximum communication latencies of 120ms. This performance is substantially better than that of threshold-based detection approaches and Kalman observer type detection approaches. These results indicate that the proposed algorithm can provide an effective, efficient and practical means of improving the cyber-resilience of IIoT control systems.

Keywords: Industrial Internet of Things; Networked Control Systems; Cyber-Attack Detection; False Data Injection; Denial-of-Service; Resilient Control; Cyber-Physical Systems.