An Integrated Methodology for Wavelet Filtering of a Pulse Wave Signal

The problem of digital filtering of a pulse wave signal relevant in cardiological diagnostics is considered, which is affected by various physiological interferences, such as baseline wander and motion artifacts. A integrated method of wavelet filtering of the pulse wave signal has been developed, which makes it possible to eliminate baseline wander and motion artifacts that distort the shape of the biosignal. The proposed technique is based on multiscale wavelet decomposition of a biosignal in terms of orthogonal Daubechies wavelets. The technique includes sequential procedures for digital processing of the pulse wave: multiscale wavelet transform; modification of the detail coefficients of the wavelet decomposition based on thresholding; reconstruction of the pulse wave signal based on original approximation coefficients and modified detailing coefficients using the inverse wavelet transform. A comparative analysis of the proposed methodology and existing approaches to filtering pulse waves (moving average filtering, median and bandpass filtering) is carried out. To obtain quantitative characteristics of the filtration efficiency assessment, simulation modeling of a pulse wave with noise of various intensity and nature was used. The high quality of pulse wave filtering using the developed technique based on multiscale wavelet transforms can serve as a reliable basis for the development of highly efficient algorithms and hardware and software systems for cardiology diagnostics.