Modulation Classification Based on Cyclic Spectral Features for Co-Channel Time-Frequency Overlapped Two-Signal

A modulation classification algorithm based on cyclic spectral features is presented to solve the problems in modulation classification for time-frequency overlapped two-signal in co-channel. Each signalpsilas features from the cyclic spectrum are extracted, and a multi-pole classifier is proposed to classify the time-frequency overlapped signals respectively in the combination of BPSK, QPSK, OQPSK and MSK signals automatically. Both theoretical analysis and simulation via Matlab verify that the algorithm is reasonable and efficient.