Application of the improved generalized demodulation time-frequency analysis method to multi-component signal decomposition

The generalized demodulation time-frequency analysis approach is a novel signal processing method, which is suitable for processing non-stationary signals, especially for processing amplitude-modulated and frequency-modulated (AM-FM) signals. The novelty of this method is that the signal whose time-frequency distributions are curved paths can be transformed into that whose time-frequency distributions are linear paths paralleling time axis by using generalized demodulation. Furthermore, the generalized demodulated signal can be decomposed into a number of mono-component signals whose instantaneous frequencies have physical meaning through the use of the maximal overlap discrete wavelet packet transform (MODWPT). Therefore, the complete time-frequency distributions of the original signal can be obtained. However, the generalized demodulation time-frequency analysis method fails to provide the time domain waveform of each mono-component signal. In this paper, aiming at the multicomponent AM-FM signal, each mono-component signal of which only has single-sided spectra, the improvement to the generalized demodulation time-frequency analysis approach has been made and the improved method is introduced into certain multicomponent AM-FM signal decomposition. The analysis results demonstrate that the validity of the proposed approach. In addition, the estimation of the phase function that is key step in the generalized demodulation time-frequency analysis approach is also discussed. (c) 2009 Elsevier B.V. All rights reserved.