Jamming Recognition Algorithm Based on Variational Mode Decomposition

Aiming to address the issue of deception jamming generated by digital radio frequency memories (DRFM), this study proposes a feature extraction algorithm based on variational mode decomposition (VMD) for deception jamming recognition and composite deception jamming recognition. Firstly, models are constructed for the real echo (RE) and the deception jamming signals. Subsequently, the variational mode decomposition is conducted. Thirdly, the features are extracted from the decomposed intrinsic mode function (IMF) and fed into the support vector machine (SVM) for classification and recognition. To mitigate the challenge of high dimensionality and reduce the complexity of the learning task, mode selection and interclass divisibility of feature selection methods are employed. The effectiveness of the proposed algorithm is verified through simulations. Prior to feature selection, a signal-to-noise ratio (SNR) of 0 dB results in a jamming recognition accuracy exceeding 95%. After feature selection, the recognition accuracy remains largely unchanged, while the recognition speed significantly improves. Compared with other methods in the same field, the recognition accuracy shows a notable improvement. Furthermore, the proposed method is evaluated for its effectiveness in recognizing composite deception jamming, and simulation results validate its performance.