Leveraging Software-Defined Radio and Deep Learning for Enhanced Physical Layer Spoofing Detection in C-V2X Communications

Vehicle-to-Everything (V2X) communication is foundational for next-generation intelligent transportation systems. However, the reliance on wireless communication exposes vehicles to significant cybersecurity threats. This paper proposes a novel framework for detecting physical layer spoofing attacks in Cellular V2X (C-V2X) networks by leveraging the flexibility of Software-Defined Radio (SDR) and the pattern recognition capabilities of Deep Learning (DL). The proposed approach is based on a graph-based representation of the signal and a Graph Neural Network (GNN) that classifies these graphs. Evaluation through simulation demonstrates that the proposed SDR-DL approach achieves high detection accuracy and robustness.