Automatic Modulation Recognition for Secondary Modulated Signals

Automatic modulation recognition (AMR) for secondary modulated signals, remains as a bottleneck, since both analog and digital modulation are adopted in one signal. To address this issue, we focus on the original signal spectrum and the 2nd order spectrum of the pre-demodulated signal, which imply modulation features. Based on this idea, a novel statistic for discrimination is proposed and the feature vector is formed. Finally, the support vector machine (SVM) is adopted to implement the classification work.