Machine Learning Systems for Connected and Autonomous Vehicle (CAV) Synthesized Datasets

This paper presents on-going research on connected and autonomous vehicle (CAV) security. The over-arching goal of the research is to secure the vehicle-to-everything (V2X) communication systems. While a previous study focused on synthesizing both normal and malicious Basic Safety Message (BSM) CoreData datasets, this research extends that previous study and explores the feasibility of applying machine learning (ML) techniques to classify the synthetized datasets. The empirical study applies four ML models: Random Forest, Extra Trees, AdaBoost, and Gradient Boosting and provides insights into their respective performances. The paper concludes with practical recommendations for future research extensions.