A Survey on Attack Detection in Networked Control Systems: Knowledge-Based and Data-Driven Approaches

Networked control systems (NCSs) integrate sensors, computation, control, and networking into physical processes and information systems to form a highly complex system. However, NCSs are susceptible to the threat of cyber-attacks because they rely on open communication networks for data transmission. At the same time, NCSs have been widely used in the field of critical infrastructures, and their security is directly related to the stable operation of the system and socio-economic development. Therefore, improving the security and robustness of NCSs has become a key challenge. In this paper, we systematically review the mainstream attack detection methods in NCSs and classify them into knowledge-based and data-driven methods based on information sources. Within this framework, the paper analyses the detection principles of common attacks, existing challenges, and corresponding research progress. In addition, combining the advantages and limitations of the two types of methods, this paper discusses the core challenges and potential research directions in the field of attack detection in NCSs.