A novel method using GA-based Clustering and spectral features for modulation classification

Because of rapid growing of radio communication technology of late years, importance of monitoring of radio waves is rising increasingly. Automatic radio signal types recognition is an important topic for both the civil and military domains. This paper proposes a high efficient technique for recognition of seven digital modulations. This technique is a pattern recognition approach. In this technique we have used the spectral characteristics for extraction the efficient features. A reduced set of parameters is derived from these coefficients and used as input to a GA-Clustering technique. The simulation results show that the proposed algorithm has high recognition accuracy to discriminate the considered radio signals.