Vehicular Network Security Against RF Jamming: An LSTM Detection System

The evolution of Intelligent Transportation Systems has ushered in a new era, offering extensive safety and comfort services through Vehicular Networks (VNs). With connected vehicles generating and disseminating vast amounts of data, encompassing critical safety and traffic information, the vulner-ability of VNs to Radio Frequency (RF) jamming has emerged as a significant concern, specially for intrusion detection systems. The dynamic nature of high-mobility vehicles and densely populated areas, coupled with the presence of roadside units, poses challenges in distinguishing between malicious attackers and legitimate transmitters. In this paper, we introduce an intelligent strategy for the classification of RF jamming attacks in VNs where Long Short-Term Memory (LSTM) networks are applied to a well-curated dataset to construct a model capable of identifying potential attacks. Various parameters, including signal characteristics, temporal aspects, and relative speeds, are considered to comprehensively address the classification task. Our proposed model demonstrates outstanding performance across multiple evaluation metrics, including accuracy, precision, recall, and F1-score. Through iterative analyses, we illustrate the incremental improvements achieved by our model compared to existing methodologies in terms of effectively classifying RF jamming in vehicular networks.