Automatic Classification of MPSK Signals Using Statistical Moments

Abstract

In this paper, an automatic classification algorithm for MPSK signals is proposed. In this algorithm, it is assumed that there is prior information about the received signal to be an MPSK type. The concept of the proposed classification algorithm is based on evaluating the statistical moments of the instantaneous phase of the received signal and using it as a key feature to classify the MPSK signals. The proposed algorithm comprises three main steps: 1) estimation of the instantaneous phase of the received signal, 2) computation of statistical moments of the estimated phase, and 3) decision about the number of phase states of the intercepted signal. The performance of the proposed algorithm is measured in terms of the success rate of classification using computer simulations. It is found that all digital modulation types of interest have been correctly classified with a success rate > 91 % at signal-to-noise ratio, SNR, of 10 dB.