A robust higher-order cyclic cumulants feature-based vector for QAM classification

The aim of this paper is to investigate performance of quadrature amplitude modulation (QAM) signals with higher order cyclic cumulants (CCs) under probability of correct classification (Pcc) measurement within a pattern recognition framework. The theoretical and simulation results are shown by choosing the feature vector, a combination of fourth and sixth order CCs. A comparison is also made with Pcc values by choosing fourth order CC alone. More importantly, the combination feature vector we have selected is robust to phase noise (PN) influence and carrier frequency offset.