Cyclic spectral features based modulation recognition

Modulation recognition of an intercepted communication signal is a fundamental problem of electromagnetic signal monitoring task arising in many fields, such as electronic surveillance and broadcasting control. A cyclic spectral features based neural network modulation recognition method is proposed. Because of the use of cyclic spectral features and the application of neural network classifier, the proposed method can efficiently recognize almost all currently used modulation types. Some computer simulation results are also reported.