Cyber Risk Analysis of Industrial Control Systems

Cyber-physical systems (CPS) represent an advanced integration of physical processes, computational resources, and communication capabilities. They are widely utilized in critical sectors such as mobility, healthcare, energy, and manufacturing. CPS depend on network communication, which creates potential vulnerabilities that can be exploited by attackers.

This work examines the difficulties associated with the protection of CPS from cyber threats, with a particular emphasis on advanced forms of stealthy cyber attack, including replay, covert, and zero dynamics attacks. Additionally, it is shown that protecting only part of the signals can increase the security of the system. A security index and its application are introduced. It is argued that the mirror attack poses an even greater threat than replay or covert attacks. The use of tools such as the extended Kalman filter allows attackers to remain stealthy even when they have incomplete knowledge of the system model. Demonstrations on a three-tank system illustrate the significant risks associated with mirror attacks and highlight the urgent need for robust detection and resilient system designs.

In order to address these challenges, an encryption scheme based on homomorphic encryption is used to secure remote monitoring. This encryption method secures not only the communication channels but also the monitoring logic, allowing computations be performed on encrypted data without exposing system parameters or signals. The effectiveness of this approach in enhancing security against cyber attacks while managing computational constraints has been demonstrated through simulations on a quadruple tank system.