Detection of Actuator Enablement Attacks by Time Petri Nets in Supervisory Control

Abstract

The escalating threat of cyber-attacks on supervisory control systems, particularly through actuator enablement manipulation, poses critical challenges to conventional security defenses. This paper presents a novel approach for detecting actuator enablement attacks based on a time Petri net. To this end, a method is presented to precisely model the system subject to actuator enablement attacks by exploiting a modified state class graph, which captures the effects of these attacks on system behavior. Then, a time-dependent defense strategy is proposed to prevent systems from suffering damage caused by actuator enablement attacks, where such damage is predefined as the system entering an unsafe state. Subsequently, the concept of “actuator enablement-safe controllability” is proposed to quantify the system’s resilience, providing a theoretical foundation for security analysis. Finally, diagnoser and verifier algorithms are developed to evaluate system safety in time-sensitive scenarios.