Biotelemetric system of centralized multi-parameter express diagnostics and monitoring of human functional state

The ideology of spatially distributed biotelemetric system of collection, processing and storage of biomedical information using the concept of Internet of Things (IoT) is developed. The system consists of personal measuring devices for registration of a set of basic vital biosignals parameters (electrocardiography (ECG), electroencephalography (EEG) photoplethysmography (PPG), impedance), multichannel central server system. A cloud service for storing and analyzing the results of measurements of biosignal parameters, algorithms and software for analyzing biosignals to select diagnostically significant parameters for assessing the human functional state on the basis of electrocardiography and electroencephalography of high resolution, heart rate and impedance parameters is developed. Approaches to the use of machine learning to determine anomalies of the functional state are analyzed and developed.
Using the methods of machine learning, the informative features for the detection of pathological conditions of the organism are determined and compared, and the classification methods that provide the highest accuracy for this task are selected. Biological signals in the time domain, frequency domain, spectral temporal and wavelet characteristics are considered. With these sets of features, the results of the work of classifiers based on decision trees, discriminant analysis, logistic regression, the method of k-nearest neighbors are obtained. Sets of features and models of machine learning are proposed, which provide the highest accuracy of recognition of the norm and pathology of the functional state of man.