Automatic modulation classification using cumulants with repeated classification attempts

In this paper, possibility for performance improvement of AMC using repeated classification attempts is analyzed. Well known AMC algorithm using fourth-order cumulants with estimation of channel information was used. Several possible schemes with combining of independent decisions or cumulant estimates derived with repeated classification attempts were proposed and analyzed. Solutions were evaluated, in terms of average probability of correct classification, using Monte-Carlo trials for different propagation environments.