A Neural Network-based Approach for Detection of Time Delay Switch Attack on Networked Control Systems

Recent improvements in communications have led to the proliferation of networked control systems (NCSs). NCSs are used to increase the efficiency and reliability of these systems. However, these connectivities between agents, sensors, and centralized controllers expose NCSs to a range of faults, failures, delays, and cyber-attacks. Detection of intelligent and new types of attacks such as time delay switched (TDS) attacks is a crucial task in the design of NCSs. The injection of the TDS attack to NCSs has the potential to provoke inefficiency or even cause instability in these systems. The TDS or delay signal attack influences the system by introducing a random delay in the process of transmitting and receiving packets. Due to the connectivity between agents, especially in distributed power grids, the injection of a TDS attack on one agent can propagate to others, which can lead to catastrophic consequences. This paper uses a neural networked-based detection algorithm to estimate the TDS attack in real-time. The performance of the proposed TDS attack detection and estimation are evaluated through simulation for two area power systems.