Supervisor Fortification Against Covert Actuator Attacks
Ruochen Tai, Liyong Lin, Rong Su · 2023
This work considers the supervisor fortification problem against covert actuator attacks. A supervisor$S^{\prime}$is said to fortify the supervisor$S$, the latter of which is non-resilient against covert actuator attacks, if$S^{\prime}$satisfies two conditions: 1) any covert actuator attack cannot cause damage infliction against$S^{\prime}$, and 2)$S^{\prime}$is control equivalent to$S$. The key result of this work is that we show the problem of determining the existence of a fortified supervisor to defend against any covert actuator attack, an “exist-for all” decidability question, is decidable. To show the decidability result, we provide a complete and sound procedure that ensures to synthesize a fortified supervisor as long as there exists one.