Automatic digital modulation recognition using spectral and statistical features with multi-layer perceptrons
M. L. Dennis Wong, Asoke Kumar Nandi · 2002
Automatic recognition of digitally modulated signals has seen increasing demand nowadays. The use of ANNs for this purpose has been popular since the late 90's. We propose an MLP recogniser implementation with better generalisation as well as an addition of a new statistical cumulant based features set for recognising ten different modulation types. Simulations show satisfactory results even with low SNR value, e.g. 98% success rate at 0 dB SNR.