On Jamming Detection: A Unified Method for Real-Time Detection Across Multiple Protocols & Attack Types

Jamming attacks continue to pose a significant threat to next-generation wireless networks, including Beyond 5G (B5G) and 6 G, by disrupting communication through intentional interference. In this paper, we introduce JAmming detection method baSed on Modulation scheme IdeNtification and Outlier Detector of un-jammed data (JASMIN), a novel jamming detection method designed to operate effectively across a broad range of network protocols and jamming scenarios. JASMIN relies solely on unjammed data during its training phase and utilizes two primary components in its detection phase: a Modulation Scheme Identification (MSI) model that classifies the legitimate signal's modulation format, and an Outlier Detector (OD) that quantifies channel noise. By comparing the predicted modulation scheme over multiple time windows with the OD's measurements of noise, JASMIN identifies abnormal interference that indicates the presence of jamming-regardless of the specific jamming strategy (e.g., constant, periodic, reactive). We demonstrate the efficacy of JASMIN on an SDR-based testbed implementing an IEEE 802.11p (V2X) communication network, employing three USRP B210 devices operating at 5.9 GHz. Evaluation results show an overall accuracy of 99.92 % under a wide range of SNR levels. Additionally, JASMIN's real-time compatibility and minimal computational overhead make it a compelling solution for modern wireless systems. To foster further innovation, we publicly release the dataset utilized in our experiments.