New structures for multirate adaptive filtering: application to interference cancelling in biomedical engineering

Presents a new structure for multirate adaptive filtering. It is based on a new mutual wavelet packets decomposition scheme. The use of this mutual wavelet packets decomposition enables the partitioning of the frequency axis into non-uniform subbands adapted to the spectral content of the signals to be filtered. A new appropriate criterion maximizing the cross-correlation between the signals is proposed for this purpose. Simulations on pulmonary microvascular pressure transients illustrate the improvement that can be achieved in filtering performance