Karplyuk Y. Increasing the accuracy of the registration of parameters for electrocardiographic diagnostics

The thesis is devoted to creating tools to improve the accuracy of registration parameters for the diagnosis of electrocardiograms, which can identify and explore the low-amplitude components of ECG, which are not available for the analysis of classical methods of ECG diagnosis. The multicriteria classification of early diagnosis systems is proposed and on that base a new system for analyzing high-resolution cardio signals is developed and proved. For the first time developed a functional-schematic model of channel ECG, which would allow the parameters optimization and to achieve an effective 12 bits at a frequency of 8 kHz, which significantly exceeds the effective resolution of existing ECG. The structure of the analog part, based on multi-channel delta-sigma ADC, is validated and developed, which significantly simplify the analog processing and improve the analog-digital conversion effective resolution up to 18-22 bits. The specialized mains adaptive filter for ECG is proposed, through which first obtained the coefficient of interference suppression of 90 dB. An algorithm analyzes the ECG signal subspace eigenvectors to obtain the diagnostic parameters is developed, which allow to analyze the fine structure of the ECG signal. The functional open software environment analysis system ECG is developed, which is based on the principle of modularity, easy scalability, unified approach to recordkeeping, the possibility of expanding the functions of analysis.