Source separation based modulation recognition of cochannel signals

In this paper, a new modulation recognition method of multiple cochannel signals is proposed. The basic idea of this method is that the null-steering beamforming technique is employed to separate each modulated signal from a ULA output data, and then a modulation recognizer, using an instantaneous frequency feature extracted by AR modeling and a RBF neural network classifier, is followed to identify the modulation type individually. Performance of the new method is assessed via computer simulations.