NOISE-RESISTANCE AND ACCURACY ASSESSMENTS ECG SIGNAL PROCESSING SYSTEMS BASED ON FILTERING METHODS

Abstract

Background. The accuracy and noise-resistance of an ECG signal processing system is influenced by many factors. But narrow-band low-frequency and broadband electrical interference have a significant impact when recording ECG signals. This interference leads to distortion of the information sections of the signal and a decrease in the accuracy of signal processing. To compensate for interference and increase the accuracy and noise-resistance of the ECG signal processing system, filtering methods are used, where filters are considered as input and output elements of the cardiographic data processing system. Aim of the research – analysis of new methods of polynomial filtering of noisy ECG signals distorted by low-frequency and high-frequency additive noise, as well as quantitative assessment of the accuracy and noise-resistance of the processing system. Materials and methods. Existing methods of filtering under the influence of electrical noise that varies over a wide frequency range of the ECG signal do not allow increasing noise-resistance and accuracy processing systems. Obtaining new filtering methods that are more accurate and resistant to interference compared to known analogues will improve the efficiency of ECG signal processing. Results. The results of the implementation of methods of polynomial digital filtering of low-frequency and high-frequency interference are presented in relation to the problem of processing ECG signals using approximating polynomial models. A computer simulation was performed to illustrate the identity of the filtered and original ECG signal forms when using the bidirectional processing procedure with the proposed filters in comparison with known filters. Quantitative estimates obtained by using test ECG signals and interference to assess the effectiveness of processing at the output of the filtering system is presented. Conclusion. The results obtained showed that filtering methods based on approximating Newton polynomials reduce the filtering error values (increase accuracy) and increase the signal-to-noise ratio (increases noise-resistance) of the ECG signal processing system when compared with known filters.

Keywords

• ecg signal processing
• filtering
• interference
• polynomial digital filters
• signal-to-noise ratio
• bi-directional filtering