Fractionally spaced decision feedback multilayer perceptron for adaptive MQAM digital mobile radio reception

The authors describe a fractionally spaced decision feedback multilayer perceptron (FSDFMLP) for adaptive multilevel quadrature amplitude modulation digital mobile radio reception that can overcome the performance degradations due to multipath fading of channels and reject cochannel interference and additive white Gaussian noise simultaneously. The two-ray model is used for the frequency selective fading channel. A fast adaptive learning algorithm called the mixed gradient based fast learning algorithm with variable learning gain and selective updates was used to train the FSDFMLP. The preliminary experimental results obtained are satisfactory, which shows that the FSDFMLP can perform more efficiently than the conventional least mean square based decision feedback filter in the presence of multipath fading of channels with non-Gaussian interferences