Higher-order statistics based modulation classification using hierarchical approach

Hierarchical based digital modulation classifier is designed using feature-extraction based method with AWGN channel. A characteristic parameter of the received information samples is used to separate between amplitude and angular modulated signals. M-ary ASK signals are separated using instantaneous amplitude of the received samples. Combination of higher-order cumulants up to order eight are computed to classify between M-ary PSK modulated signals. A new feature is proposed in the decision tree of the classifier to separate QPSK and 8PSK modulation. Simulation results are used to verify that this approach is robust under low SNR.