Quasi-Haar wavelet and modulation identification of digital signals

A quasi-Haar wavelet that has better localization characteristics in the frequency domain than the Haar wavelet is defined. Furthermore, the feasibility of using the sinusoidal quasi-Haar wavelet to classify MPSK and MFSK is demonstrated, and the relationship between the amplitude of wavelet coefficients and frequency or phase variance is deduced in detail. Finally, simulation results are given. Both the theoretical analysis and the simulation results show that the sinusoidal quasi-Haar wavelet not only can be used for modulation identification of communication signals, but also has better anti-noise performance under the appropriate SNR.