The relationship between Haar wavelet decomposition coefficients and modulated signal parameters is discussed. A new modulation classification method is presented. The new method uses the amplitude,frequency and phase information derived from Haar wavelet decomposition as feature vectors to distinguish the modulation types of M-ary Frequency-Shift Keying (MFSK), M-ary Phase-Shift Keying (MPSK) and Quadrature Amplitude Modulation (QAM) modulation types. A parallel combined classifier is designed based on these feature vectors. The overall successful recognition rate of 92.4% can be achieved even at a low Signal-to-Noise Ratio (SNR) of 5dB.
展开▼
