A Stochastic Bayesian Game for Securing Secondary Frequency Control of Microgrids Against Spoofing Attacks With Incomplete Information

While wireless communication has been implemented for the data exchange in the secondary frequency control of microgrids, the wireless links also open doors to spoofing attacks. Most existing game-theoretic approaches on securing control systems of microgrids against wireless spoofing attacks assume complete information. However, the defense scheme under perfect information assumption could lead to severe resource waste and significant detection delay due to the high over-defense rate. In this paper, we design a defense policy generation method for securing microgrids secondary frequency control facing spoofing attacks and incomplete observation. We formulate a multi-stage two-player stochastic Bayesian game (SBG) when the identity of the defender's opponent is uncertain. Furthermore, we propose a posterior identity belief update method, where Bayesian Nash equilibrium (NE) is considered to derive the boundary identity belief. Under the proposed SBG framework, an identity-dependent optimal defense scheme is obtained to simultaneously secure microgrids against potential spoofing attacks and reduce the over-defense rate. Comparison studies show that the proposed SBG-based defense policy can improve defense performance and significantly reduce over-defense rates.