MATHEMATICAL, ALGORITHMICAL AND SOFTWARE SUPPORT FOR WAVELET PROCESSING OF PULSE SIGNALS IN THE MORLET BASIS

Abstract

 The work developed a method based on wavelet transform in the Morlet basis and created the appropriate algorithmic and software for processing pulse signals in order to detect its structural fluctuations in time at different scales of observation. It was found that the use of wavelet transform allows for a detailed analysis of the time-frequency structure of the pulse signal, which, in turn, provides the possibility of detecting its structural fluctuations at different time scales - an important aspect in diagnosing changes in the state of the human vascular system. The research conducted in the work provided the calculation of new diagnostic features regarding the state of human vessels in the form of 2D and 3D spectral images of wavelet coefficients, which increased the level of diagnostic capabilities of computer photoplethysmographs.
 To register pulse signal data as objects of wavelet processing, a computer photoplethysmograph developed on the basis of the Department of Biotechnical Means of TNTU was used. As a sensor for registering the pulse signal, an optical sensor was used, which includes an infrared LED with a radiation range of 625–740 nm, which acts as a light source, and a phototransistor that records the signal reflected from the vascular surface. Tools from the Matlab environment were utilized to develop the software for pulse signal processing, with the Matlab Guide module employed to implement automated processing within computer-based photoplethysmographic systems. 

 The efficiency of the developed software for processing the pulse signal was studied in order to detect timely changes in the state of the vascular system. The results showed that the developed software functions correctly and reliably reflects vascular changes, which is confirmed in the form of 3D and averaged 2D spectral images of wavelet coefficients, which act as new diagnostic indicators.