Signature and anomaly based intrusion detection system for secure IoTs and V2G communication

Abstract

Cybersecurity is considered a top priority across all organizations, from corporations to governmental agencies. Every year, we hear about major cyber security breaches in many organizations, especially targeting smart grids with the Internet of Things (IoT) and electric vehicles (EVs) and hindering their operation. Cybersecurity attacks are multifaceted and may initiate with identifying a vulnerability, exploiting that vulnerability to gain shell access, elevating privileges, performing lateral movement, executing commands, exfiltrating data, covering tracks, and keeping persistent access. Based on a comprehensive literature review and analysis, a joint detection system using signature-based and anomaly-based intrusion detection systems (SBaIDS) that utilizes different machine learning algorithms based on a hybrid model is proposed. The novel system effectively detects cybersecurity attacks and performs non-traditional detection mechanisms for IoT and vehicle-to-grid (V2G) communication. The implementation uses the enhanced dataset and presents the result in terms of accuracy, precision, recall, and F1 Score given each attack scenario. The proposed model’s performance shows good results compared to the previous work that uses a Support Vector Machine (SVM) running SPARK. The detection ratio is found to be more than 96%, compared to recent work on a hybrid intrusion detection system.