Two-Layer Deception Model Based on Signaling Games Against Cyber Attacks on Cyber-Physical Systems

Cyber-physical systems (CPS) are increasingly vulnerable to sophisticated cyber-attacks that can target multiple layers within the system. To strengthen defenses against these complex threats, deception-based techniques have emerged as a promising solution. While previous research has primarily focused on single-layer deception strategies, the authors argue that a multi-layer approach is essential for effectively countering advanced attackers capable of perceiving information across both the application and network layers. In this work, we propose a two-layer deception model based on signaling games to enhance the defense of CPS. Our model captures the dynamic, non-cooperative interactions between the attacker and defender under conditions of incomplete information. Unlike existing approaches, our model expands the defender's action space to incorporate deception at both the application and network layers, while maintaining the attacker's uncertainty about the true system type. Through analytical and simulation results, we identify the Perfect Bayesian Nash Equilibrium (PBNE) strategies for both players. Our findings demonstrate that the two-layer deception model significantly outperforms single-layer deception in deceiving the attacker and improving system resilience, particularly against sophisticated adversaries capable of perceiving information across multiple layers.