Automatic classification of communication signals using higher order statistics

A method for the automatic classification of digitally modulated communication signals is introduced. Different nonlinearities applied to the complex envelope of the signal under classification are used to exploit differences in the higher-order moment spaces of the time-discrete modulating process. These differences manifest themselves in lines in the spectrum of the transformed signal, associated with the unknown carrier frequency and baud rate. The lines are detected by the periodogram analysis. Their existence, position and amplitude constitute a robust feature for separating 2ASK, 2PSK, 4PSK, MSK and 2FSK. A complete statistical analysis of the classification performance in terms of the probability of detection and the false alarm rate is carried out and compared with simulated data. The obtained results are found to be valid for a wide range of modulation parameters