Wavelet Analysis And Extraction Of The Fine Features Of Typical Communication Signals

Wavelet analysis is a rising theory, which has extensive applications in different scientific and technological fields. After an introduction to the concept and property of the wavelet and present applications in communications, this dissertation focuses the research on the wavelet's application to extraction and analysis of the fine features of typical communication signals, and four problems are discussed:(1) An new method for the feature extraction and pattern recognition of typical communication signals A distinct method based on wavelet transform is proposed. It can extract the information of amplitude, frequency and phase of communication signal effectively. The inter-class and intra-class classification of ASK, FSK and PSK is made.(2) Zero-cross feature of MPSK signals'wavelet transfom When MPSK signals are analyzed with wavelet transform, a kind of novel zero-cross phenomenon is found and illustrated in graphics. The relationship between the scales at zero-cross areas and the mode of phase difference is developed.(3) Frequency offset on wavelet packet transform The relationship of position of impulse response h[n] and g[n] of wavelet packet filters in sequence of frequency is derived, and the physical significance of the frequency offset is discussed from both Mallat algorithm and algorithm a 'trous respectively.(4) Periodicity and pseudo-modulation when sampling sinusoidal Signals The periodicity and pseudo-modulation due to non-integral sampling of sinusoidal signal are analyzed.