The identification of modulation type is an important topic in signal detection. It has seen increasing demand over recent years. Most of identifiers (techniques) have been proposed for additive white Gaussian noise (AWGN) environments. However, communication environments such as wireless communication channels suffer from dispersion. In this paper we present a high efficient method (HEDMI) to automatically characterize and classify digital modulation types in fading environment. In HEDMI, channel effects are mitigated using an equalizer. Fourth, sixth and eighth orders of statistical moments and cumulants are used as features and a multiclass SVM based classifier, recognizes the type of digital modulations based on extracted features. Experimental results indicate that HEDMI has high efficiency and robust performance to discriminate the different kinds of modulation in fading environment even at low SNRs.

Keywords