Algorithm for automatic recognition of PSK and QAM with unique classifier based on features and threshold levels

In this paper, we present a unique modulation classification method that is based on determining an attractive relation between higher-order cumulants (HOCs) using a decision tree-classifier to improve the extracted features employed for the recognition of modulation schemes, such as phase shift keying (PSK) and quadrature amplitude modulation (QAM). A threshold algorithm is applied to the proposed classifier, which consists of sub-classifiers, each comprising a single feature, and each being capable of distinguishing the modulation types individually. In this work, a high-accuracy classifier system is utilized to recognize modulation schemes, such as QAM (16, 32, 64, 128, and 256) and (2, 4, and 8) PSK at a low signal-to-noise ratio (SNR). In this study, 1000 signals are studied for each SNR of –5 dB to 30 dB. The most prominent results of the classifier decisions range from 88% to 100% with regard to distinguishing the same types of PSK and QAM. In the long run, the proposed classifier module will be advantageous in terms of accuracy and computational complexity relative to the other classifiers in the literature. The results demonstrate that the proposed algorithm has a significantly better classification accuracy in comparison with the previously proposed ones. •A new method for automatic modulation classification is presented.•Optimize the extracted features distribution using logarithmic properties.•Multiple logarithmic hierarchical classifier outperforms over conventional classifiers.•Enhanced classification performance through multiple tree algorithm.•Simulation results confirmed the efficiency of the proposed algorithm under different noise scenarios.