A Modulation Recognition Algorithm based on Cyclic Spectrum and SVM Classification

Abstract

The modulation recognition algorithm identifies the modulation mode of the communication signal to determine the signal type contained in the communication signal. A modulation recognition algorithm based on cyclic spectrum and support vector machine classification is designed. The algorithm includes two aspects: feature extraction and classifier design. By selecting the second-order cyclic spectrum feature of signal with high separability and strong stability, and the statistical moment feature of signal amplitude realized in frequency domain, the algorithm is realized by using support vector machine classifier according to the selected feature. Finally, the experiment simulation is carried out.