Advanced methods for digital quadrature and offset modulation classification

The authors address the problem of constructing classifiers of digital quadrature and offset modulated signals by the likelihood approach and the M^th-law approach. In the first approach, they start from the likelihood functionals (LF) of such signals in additive white Gaussian noise and, based on these LFs, derive easily implementable classifiers. In the second approach, M^th-law classifiers are examined for classifying such signals. It is shown that certain versions of the LF classifier are closely related to the M^th-law classifier in terms of both implementation and performance, at least in the low signal-to-noise ratio regime