WCGAN-Based Cyber-Attacks Detection System in the EV Charging Infrastructure

Transportation electrification and connected automated vehicle infrastructures are the future of sustainable e-mobility, fueling the robust deployment of vehicle charging infrastructures and evolving network and communication architectures. The incumbent technologies of sustainable e-mobility migrate the inherent vulnerabilities in software, hardware, protocols, communication, and human that state-funded or ill-willed cyber attackers could exploit. The existing deep learning-based detection algorithms suffer from constrained performance due to insufficient cyberattack data. Inspired by the synthetic data generation by the generative adversarial network (GAN), we propose the external classifier Wasserstein condition GAN (EC-WCGAN)-based network intrusion detection systems (NIDS) to detect the distributed denial of service (DDoS) attacks in the EV Charging infrastructures. Moreover, it can inform the different DDoS attack classes as well. The proposed method surpasses the DL-based model in terms of data usage and performance metrics. This method enhances the detection by generating plausible synthetic data for the low sample classes and training the classifier to get more than 99% performance metrics.