A Generative Adversarial Network Based Defense Strategy Against Stealthy False Data Injection Attacks Targeting Prices in V2G System

With the remarkable growth of electric vehicles (EVs), the spotlight has been firmly shone on vehicle-to-grid (V2G) technologies. In the context of V2G systems, concealed False Data Injection Attacks (FDIA) targeting the dispatch prices of EVs possess the potential to severely upset the delicate supply-demand equilibrium of the power grid. Defending against stealthy FDIAs presents challenges because it requires anticipating the model used by the attacker. In this context, this paper proposed a defense strategy without the requirement of priori attack knowledge. The proposed method uses a generative adversarial network (GAN) to generate stealthy FDIA vectors to improve the successful detection rate of attacks. The experimental results show that this strategy achieves FDIA defense under unknown attacker models, reduces the impact of FDIAs on social welfare, and enhances the power grid’s resilience to FDIAs.