A synchrosqueezed wavelet transform enhanced by extended analytical mode decomposition method for dynamic signal reconstruction

摘要

A synchrosqueezed wavelet transform combined by wavelet analysis and time-frequency reallocation method was recently developed to improve the quality of time-frequency representation and construct the components of a signal. However, the quality of this transform highly depends on the parameters of wavelet functions. To overcome the effects of wavelet function parameters on signal decomposition, an analytical mode decomposition method was proposed to reconstruct each component. The analytical mode decomposition selects the bisecting frequency between any two adjacent instantaneous frequencies estimated from a preliminary synchrosqueezed wavelet analysis and analytically extracts the individual components from a signal between any two bisecting frequencies. Two examples including a frequency modulated amplitude decayed signal, and seismic responses of a two-story building are analyzed. It is demonstrated that the analytically derived components are insensitive to the selection parameters of wavelet function and bisecting frequency. The proposed method is superior to the synchrosqueezed wavelet transform method for dynamic signal reconstruction.