Abstract: Wavelet transform based on analytic signal is proposed, with its properties discussed. By this method, the energy of the signal in a certain frequency subband, which may change with time, can be calculated out like Wigner distribution. The outcome of the transform can be visualized easily, overcoming the drawbacks of common wavelet transform. Three examples are given, which include its application on time-frequency analysis, denoising and fault diagnosis.

Key words: wavelet transform, analytic signal, fault diagnosis