Blind modulation classification in the presence of carrier frequency offset

Blind modulation classification deals with identification of modulation formats from received signal without the knowledge of the type of modulation transmitted. The problem becomes more challenging in real world scenarios when there are synchronization errors such as frequency offset and timing offset, and multi-path fading. In this paper, we consider the problem of modulation classification in the presence of carrier frequency offset. Several algorithms, based on higher-order cumulants of differentially processed received signal and cyclic cumulants of the received signal, have been recently proposed for blind classification in the presence of frequency offset. Since the variance of the estimates of these features is high, more data and/or high SNR is required for good classification performance. We consider a 9-class problem and propose a hierarchical classification using a combination of moments of the received and differentially processed received signal, unlike in where they only use the cumulants of the differentially processed received signal. The motivation for our approach comes from the fact that for a given data size, the estimates of the moments of the received signal have less variance compared to those of differentially processed received signal. Simulations are used to illustrate the performance of the proposed algorithm. The results show that our approach yields better detection probability than that of which considers 10-class problem of which 8 are common to our class.