Recognition of digital modulation signals in shortwave communications

There has been no effective way to recognize modulation signals in shortwave communications until now.The authors developed a method for digital modulation recognition based on wavelet packet transform,high order cumulants and support vector machine.In the method,the energy values and cumulants obtained from wavelet packet transform are extracted as the eigenvectors,then hierarchical classifiers based on support vector machines are used as the classifiers to modulate and recognize the signals.This hierarchical method has a higher recognition rate than other non-hierarchical ones.Classification results for 10 modulation types including FSK,PSK etc.showed that the method has higher recognition performance in low SNR environments and it has wide application in recognition of modulated signals in shortwave communications.