Machine Learning-based Secure Communication Framework for Connected Autonomous Vehicles

As connected autonomous vehicles (CAVs) gain popularity, new challenges emerge in ensuring secure network connectivity. In this study, we present a machine learning (ML)-based secure communication system for autonomous vehicles. Our main goal is to distinguish between harmful and non-malicious message transfers in order to strengthen communication channel security. We used a unique framework to bin float values to make them more suitable for classification. We then fed these values into a classifier model. The framework utilizes a dataset comprising 63 features and 450,000 records. Among various ML models evaluated, random forest (RF) emerged superior in classifying CAVs messages. The RF model, trained on this dataset, achieved an impressive accuracy of 94.26%. This underscores the potential of ML techniques, especially RF, in fortifying the security of networked autonomous vehicles. This study contributes to ongoing efforts to establish secure and reliable communication systems for CAV networks.