Y.A. Chelebaeva1, O.V. Melnik2, S.V. Chelebaev3

1–3 Ryazan State Radio Engineering University named after V.F. Utkin (Ryazan, Russia)

The task of real-time heart rate analysis is to detect early arrhythmias in order to treat them and prevent life-threatening arrhythmias. To solve the problem of classifying heart rhythm features based on cardiorhythmogram processing, an apparatus of artificial neural networks can be used. A more detailed analysis requires a comprehensive interpretation of the identified features. This is due to the fact that a single detected sign may not be the cause of the disease, but may be due to the peculiarities of the body examined, or inaccuracies of measurements. For this, a mathematical apparatus based on fuzzy logic is chosen.

The purpose is development of fuzzy inference rules for a model of complex interpretation of cardiorhythmogram signal features.

The input clear numerical values are informative features identified by the neural network structures of the cardiorhythmogram signal processing subsystem.

Analytical dependencies are proposed for fuzzification of input clear numerical values of fuzzy system of complex interpretation of detected features.

Fuzzy inference rules have been developed to calculate the resulting functions of belonging to the output value of the model. The formation of signs of normal sinus rhythm, atrial fibrillation, arrhythmia, premature ventricular contraction, atrial extrasystole is illustrated.

The output variables of the fuzzy model have been defazzified.

Analysis of compliance of the features identified by the fuzzy model with the database cardiorhythmograms annotations CU Ventricular Tachyarrythmia Database, Congestive Heart Failure RR Interval Database, MIT-BIH Atrial Fibrillation Database, Normal Sinus Rhythm RR Interval Database и CAST RR Interval Sub-Study Database was carried out https://physionet.org/cgi-bin/atm/ATM.

The results of cardiorhythmograms processing confirm the correctness of the developed models.

Different arrhythmias can exist together, and the fuzzy inference rules developed allow this combination to be detected in parallel. This approach to the construction of heart rate control systems in real time can be used both for monitoring already diagnosed cardiovascular diseases, especially in intensive care units, and for the prevention and early diagnosis of signs of arrhythmia in persons with high myocardial risk.

Chelebaeva Y.A., Melnik O.V., Chelebaev S.V. Development of fuzzy inference rules for a model of integrated interpretation of cardiorhythmogram signal features. Biomedicine Radioengineering. 2022. V. 25. № 4. Р. 87-97. DOI: https://doi.org/10.18127/j15604136-202204-11 (In Russian)

1. Bereznyy E.A., Rubin A.M., Utekhina G.A. Prakticheskaya kardioritmografiya. M.: Neo. 2005. 140 s. (in Russian).
2. Rukovodstvo po kardiologii: Uchebnoye posobiye v 3 t. / Pod red. G.I. Storozhakova, A.A. Gorbachenkova. 2009. T. 3. 512 s. (in Russian).
3. Melnik O.V., Chelebayev S.V., Chelebayeva Yu.A. Analiz serdechnogo ritma v rezhime realnogo vremeni na osnove iskusstvennykh neyronnykh setey. Biotekhnosfera. 2016. № 6. S. 33–39. (in Russian).
4. Chelebayeva Yu.A. Razrabotka nechetkoy modeli kompleksnoy interpretatsii priznakov na osnove analiza signalov kardioritmogrammy. Biomeditsinskaya radioelektronika. 2020. T. 23. № 4. S. 85–93. (in Russian).