Securing electric transportation networks: A machine learning-driven cyber threat detection

Abstract

The study examines the cybersecurity environment of electric transportation networks using a machine learning-based methodology. It analyzes the behaviors of electric vehicles, charging patterns, cyber threat occurrences, and the performance of machine learning models. An analysis of electric vehicle (EV) data shows that there are differences in battery capacity and distances covered, suggesting the presence of possible weaknesses across different cars. Cyber threat logs provide a comprehensive view of the various levels of threat severity and the time it takes to discover them, illustrating the ever-changing nature of cyber threats in the network. Machine learning models have varying performance; ML003 and ML005 exhibit excellent accuracy and precision in threat identification, whilst ML002 shows significantly lower metrics. These results highlight the need of implementing flexible cybersecurity solutions to handle different electric vehicle behaviors and effectively reduce cyber risks. This research emphasizes the need of using proactive threat detection tactics in order to effectively address high-severity attacks. It also highlights the need for ongoing improvement of machine learning models to strengthen network security. This study enhances our comprehension of cybersecurity obstacles in electric transportation networks, highlighting the crucial significance of machine learning-based analysis in strengthening network resilience against ever-changing cyber threats.