A method of quadrature amplitude modulation signals identification at low signal-to-noise ratios

Quadrature Amplitude Modulation (M-QAM) was developed to rapidly and automatically identify the modulation levels of digitally modulated signals at low signal-to-noise ratios (SNR). The method uses wavelet transforms with the optimal scale combining with manifold learning method to identify the modulation levels of the M-QAM signals. Simulation results show that when the SNR is not lower than - 22 dB, the percentage of correct identification of M-QAM is higher than 90%. The results show that the method can rapidly acquire good performance at low SNR.