Maximum likelihood PSK classification using the DFT of phase histogram

A method is presented for the classification of multilevel PSK signals which uses a maximum likelihood function on the DFT of phase histogram. An expression for this likelihood function is derived, which results in a simple function involving modified Bessel functions. The method has been evaluated for the classification of CW-8PSK and BPSK/QPSK and the error performance is comparable to or better than other methods used for comparison. However the method performs well in terms of computational complexity, which makes it attractive for the classification of PSK signals.