A Counteracting Resilience Strategy for Denial-of-Service Scenarios in Networked Control Systems

In this manuscript, we introduce a model predictive control (MPC) strategy tailored to improve the security and resilience of constrained cyber-physical systems facing attacks whose target is the communication channel. Building upon a recently developed resilient control framework, the improvement consists in addressing the challenges posed by time length varying denial-of-service attacks. Its primary focus lies in its capacity to mitigate undesired system behaviors stemming from external malicious actions.Central to our strategy consists in using set-membership conditions, facilitating swift detection of data integrity anomalies and enabling the implementation of appropriate countermeasures. By harnessing set-theoretic based model predictive control ideas, the resulting solution emphasizes its resilient nature in the efficient handling of explicit plant constraints. Moreover, the proposed scheme is designed to ensure the system's safe operation during attacks and minimizing performance degradation. To further bolster resilience, a network refresh procedure is added, which serves to securely restore attack-free communications should the initially deployed countermeasures prove insufficient. Through extensive simulations, wherein we apply our methodology to a tracking problem involving a planar mass-point system, the effectiveness of the resilient scheme here proposed is proved. These simulations confirm the capability of our control architecture to maintain constraint satisfaction across a spectrum of possible denial of service attack scenarios, thereby underlining its robustness and practical usefulness in safeguarding cyber-physical systems against malicious threats.