Hierarchical Automatic Recognition of MPSK and MQAM Signals Using LVQNN and RBFNN

In this paper, two kinds of artificial neural network have been combined, and used as recogniser to complete the hierarchical automatic recognition of MPSK and MQAM singals. They are learning vector quantization neural networks (LVQNN) and radial-basis function neural networks (RBFNN) respectively. The former is used for classes of signal, and the latter is used for subclasses. Simulations show the performance of LVQNN and RBFNN classifiers is satisfying for six different digital modulation schemes, even at signal-to-noise ratios (SNR) as low as 6 dB.