Application of adaptive kernel time-frequency distribution in cross-components suppression
Lei Fen-xia · Systems engineering and electronics · 2004
The limitations of time-frequency distribution (TFD) with fixed kernel on cross-components suppression are analyzed and the improvement of time-frequency resolution using signal-dependent adaptive kernel TFD is explored. An improved algorithm for calculating the signal-dependent adaptive kernel time-frequency distribution is put forward based on the radially-Gaussian signal dependent representation. The algorithm adopts the short-time ambiguity function (STAF) and the adaptive kernel varying with time. It can differentiate the details of multi-component signals in TFD. Simulation results show that the new algorithm possesses a high time-frequency resolution and has no interferency terms. At the same time it is applicable to analysis of long signals and on-line signal processing.