Automatic modulation recognition using RBFNN and efficient features in fading channels

Automatic modulation recognition plays an important role for network traffic administration, different data rate allocation, signal confirmation, interference identification, software radios, multi-drop networks, intelligent modems, etc. In this paper we present a high efficient method that automatically characterizes the digital modulation types in fading channels. In this recognizer, a radial basis neural network is used as the classifier. Channel effects are mitigated using an equalizer. A combination set of higher order moments and higher order cumulants up to eighth are proposed as the effective features. Experimental results indicate that the proposed identifier has high recognition accuracy at low level of the signal to noise ratios (SNRs).