Instantaneous frequency estimation using Malvar wavelet-based discrete evolutionary transform

A method - based on the discrete evolutionary transform (DET) - to estimate the instantaneous frequency (IF) of a noise corrupted nonstationary signal is presented. Computation of the evolutionary kernel employs Malvar wavelets. An expression of IF exploits a characteristic of the DET. Implementation of IF estimation is done by masking and noise cancelling. The masking enables the procedure to be valid for multi-component signals. Noise cancellation is possible by adaptive filtering - based on a normalized least mean squares algorithm - and to reduce variation of the resulting IF estimates. Examples are used to illustrate the performance for mono- and multi-component signals in noisy situations.