Automatic modulation classification using compressive sensing based on High-Order Cumulants

Abstract

High-Order Cumulants (HOCs) is widely used as the feature in automatic modulation classification (AMC) for it has the outstanding resiliency to noise. However, traditional works require more than Nyquist sampling rate for HOCs extraction. In this work, a HOCs-based method based on compressive sensing (CS-HOC) is introduced. Without reconstructing the original signal, we propose a scheme to estimate the fourth-order and sixth-order cumulants of unknown signals based on received compressive samples, which greatly reduces the number of samples. In order to deduce the sparse representation of fourth-order and sixth-order statistic, the Walsh-Hadamard Transform is brought in. From the simulations we can see that the CS-HOC method distinctly promotes the classification rate compared with traditional sampling schemes.