Registration Number
0225U005123,
(0125U002047)
, R & D reports
Title
Development of technology for objective control of functional capabilities and stress of military personnel based on miniature electrocardiographs and machine learning.
popup.stage_title
Розроблення та валідація вирішувальних правил на основі машинного навчання, апробація технології у потенційних споживачів.
Head
Chaikovskyi Illia A., Кандидат медичних наук
Registration Date
25-12-2025
Organization
V.M. Glushkov Institute of Cybernetics of the National Academy of Sciences of Ukraine
popup.description1
Development of technology for objective control of functional capabilities, combat and operational stress of personnel of the Defense Forces of Ukraine using ultra-miniature electrocardiographic devices and artificial intelligence.
popup.description2
A large array of electrocardiographic data was created, which includes a total of about 30 thousand electrocardiograms, both of military personnel in various circumstances and of civilians. The array of electrocardiographic data includes a large control group, also consisting of both military personnel and civilians. Pre-processing of all electrocardiographic records was carried out, a large array of electrocardiographic indicators was obtained. Analysis of data sets from different sources was carried out for content and compatibility, normalization of units of measurement. Informative parameters were selected for use as arguments of the machine learning model. Data sets were selected for use as training, test and validation samples.
An important result of the project was the development and validation of interrelated methods for objective control of functional capabilities, combat and operational stress of military personnel using /superminiature software and hardware electrocardiographic complexes suitable for use in field conditions.
The methods are based on the fact obtained for the first time in the course of these studies that military personnel who participated in combat operations form two very characteristic separate clusters in the multiparametric space of electrocardiographic research results.
Based on machine learning, a model and software were built to determine the main, most significant indicators of cardiac contractile function from ECG parameters, as an important factor of functional capabilities. Parameters were determined, the values of which can be predicted with the smallest average absolute and mean square errors. Also, the possibility of determining the most clinically significant EchoCG parameters from ECG parameters based on machine learning and a multi-class classification model was investigated.
Product Description
popup.authors
Sharypanov Anton V.
Andrii L. Golovynskyi
Tetiana M. Ryzhenko
Vadym G. Tulchynsky
Illya A. Chaikovsky
Anton Popov
popup.nrat_date
2025-12-25
