Abstract
Automatic Modulation classification (AMC) is the process of identifying modulation type of a detected signal without prior information. AMC is critical for Signal Intelligence (SIGINT) and communication jamming Many AMC algorithms exist in the literature targeting different set of modulations, known modulation parameters, and noise distributions. However, only a hand full investigate the effect of receiver parameter ambiguity on the AMC performance. We present an AMC algorithm that classifies 10 single carrier digital modulations in presence of uncertainties in parameters, such as, carrier frequency offset, pulse shaping, timing offset, and symbol rate. The proposed algorithm is based on Fourier Transformation and constellation diagram. A mix of Support Vector Machines (SVMs) and decision-tree is used as the classifier. A computer simulation is used to show the results of propposed AMC algorithm.