Physical-Layer Security in OFDM-Based PNC-Enabled IoT: Analysis and Detection of Symbol-Level Jamming Attacks

This paper investigates Symbol-Level Jamming (SLJ) detection in Orthogonal Frequency Division Multiplexing (OFDM)-based Two-Way Relay Channel (TWRC) systems using synchronous Physical-layer Network Coding (PNC). The system includes a relay node exposed to SLJ, characterized by Jamming-to-Signal Ratio (JSR) and duty cycle parameters. Three pilot-assisted detection methods are evaluated: energybased, Error Vector Magnitude (EVM)-based, and one-class Support Vector Machine (SVM)-based. Simulation results demonstrate that while the energy-based detection performs adequately under strong jamming, its sensitivity diminishes at lower jamming power levels. In a mobile network with low JSR levels (e.g., JSR=-5 dB ), the EVM-based detector improves robustness by leveraging constellation distortion metrics, achieving accuracy and precision of 0.97 and 0.87, respectively. On the other hand, the SVM-based method, trained on EVM features extracted from clean symbols, achieves the highest detection performance, with accuracy and precision of 0.99 and 0.97, respectively. Metric selection depends on the application— precision is critical for jammer localization, while accuracy is for network rerouting. The proposed approach enhances SLJ resilience in OFDM-PNC systems.