Our Research

Group leader -- Associate Professor, Ph.D. A. O. Popov

Work Areas:

We study a wide range of physiological signals and biomedical data, including:

• Electroencephalography (EEG) for the study of mental health, burnout, epilepsy, Parkinson's disease, and brain-computer interfaces.
• Electrocardiography (ECG) and heart rate variability (HRV) for the analysis of arrhythmias, sleep apnea, and cardiovascular risk assessment.
• Oculography and postural control signals for the detection of concussions, mild traumatic brain injury, and other neurological disorders.
• Electrical activity and action potentials of cardiac muscle cells for the assessment of cardiotoxicity and drug safety studies.
• Breath sound analysis for the detection and monitoring of respiratory diseases, such as bronchitis and COPD.

From a methodological point of view, we focus on machine learning, deep learning, fractal and large-scale analysis, information-theoretic approaches, as well as modern methods of time-frequency and nonlinear signal processing.

For more detailed information, please visit the scientific group's website.

Group leader -- professor, Dr. Sc. Yu. V. Prokopenko

Areas of work:

• quantum electron transport
• nanostructure modeling
• physics of low-dimensional systems
• solid-state physics
• solid-state electronics
• materials science
• photonics

More details on the group page.

Group leader -- O. O. Vityaz.

Directions of work:

• Theory of electronic circuits
• Structural and Parametric Identification of Equivalent Circuit Models for Two-Terminal Components
• Triangulation of Mathematical Models of Three-Terminal Components
• Circuit Modeling of Multiphysics Systems Based on Electrodynamic Analogies
• Circuit-Level Analysis of Nonlinear Electronic Circuits in the Periodic Steady-State
• Feedback Theory and Synthesis of Harmonic Signal Oscillators
• Mathematical Modeling and Identification of Diagnostic Parameters of Human Respiratory Biomechanics

The work of the laboratory is aimed at conducting the educational process and scientific research of the Department of Physical and Biomedical Electronics

Directions of work:

• Components, devices and systems of biomedical electronics.
• Methods and computer diagnostic complexes for electrocardiography, ultrasound, electroencephalography, computed tomography.
• Systems of intelligent monitoring of the state of the organism, biometric personal identification systems.
• Automated interpretation of diagnostic data and biomedical expert systems.
• Means and systems of analysis of signals of electrocardiography, ultrasound, electroencephalography and computed tomography.
• Nanobiocomponents, biomedical technologies and systems.
• Thermographic diagnostic systems.
• Modeling of biomedical systems.

More details on the laboratory page.

Mathematical models of signals characterizing the functional state of a person have been developed. For electrocardiograms and electroencephalograms, new types of integral transformations have been applied, which allow obtaining model signals that are close to real ones and minimized in the number of parameters.

For signals of hemoglobin and glucose measuring devices, it is planned to obtain optimal forms of spectral characteristics of photodetector sensitivity, which provide reliable diagnostics of diseases. Adaptive matched filters have been used when processing pulse wave signals, which allowed to increase the signal-to-noise ratio and ensure the reliability of diagnostic and physiotherapy equipment. Studies of the interaction of magnetic fields and laser radiation in physiotherapy equipment with the parameters of various human physiological systems and, accordingly, with diagnostic signals have been conducted.

A method for analyzing electrocardiograms and electroencephalograms based on the eigensubspace method was developed for the first time. This made it possible to detect late potentials of the atria and ventricles for the diagnosis of cardiac tachycardia.

An algorithm was created and procedures for spectral evaluation and identification of models of low-amplitude components of ECG signals were improved. When processing electroencephalograms using the eigenvector method, the main features of EEG signals characteristic of the early stage of epilepsy were identified. A program for their automatic
recognition and making a primary diagnosis.

In the direction of developing methods of diagnostics and physiotherapy, a laboratory sample of a closed magnetolaser physiotherapy system with feedback on the parameters of the pulse wave, electrocardiogram, and oximetric signal was created. The synthesis of the structure of this system and the analysis of the interactions of individual subsystems were performed at a level corresponding to the world. It differs qualitatively in the use of new diagnostic parameters of various signals and quantitatively in terms of the effectiveness of physiotherapy, since it uses adaptive modes of functioning in relation to a specific patient.

The idea of ​​adaptation is also implemented in the choice of signal processing technology. For this, original signal processing techniques were used, for example, wavelet transform and neural networks.

The work obtained results that are directly related to the practical application of the developed methods in diagnostic and physiotherapy equipment. Firstly, this is the development of new diagnostic parameters focused on computer information processing technologies. Secondly, this is the creation of original circuit solutions and software taking into account the modern level of computing technology and the component base of electronics. Thirdly, the improvement of magnetolaser therapy tools based on the use of biological feedback with an adapted system of diagnostic parameters.

The analysis of publications on the topic of the project and the expected results formulated above give reason to believe that the level of work performance corresponds to the world level, allows us to initiate the creation of a new generation of intellectualized physiotherapeutic devices and complexes.

The results of the research can be used in the training of specialists in biomedical electronics. Students and graduate students will be involved in this work, a number of staff will be involved, and it will be submitted for diploma design and in attestation master's theses. Separate sections of the work will be developed in the candidate theses of graduate students.

Based on the work, a guaranteed competitive method and technology of combined diagnostics and therapy with biofeedback is created. This method is promising for express diagnostics and treatment, as well as in the monitoring mode and preventive measures to control the functional state and health of the population of Ukraine. It creates opportunities for the implementation of a new technology for the treatment and diagnosis of diseases of the cardiovascular and nervous systems.

• The work performed corresponds to the current problems of medicine related to the treatment of diseases of the heart, brain and blood;
• The work was performed in several areas: electrocardiography, electroencephalography, glucometry and magnetic-laser physiotherapy, and the results obtained provide a basis for further research into their use in physiotherapeutic equipment;
• the work is characterized by scientific novelty in approaches to specific tasks of diagnosing diseases such as cardiac tachyarrhythmia, epilepsy, diabetes, as well as a closed physiotherapy system with biological feedback;
• the results of the research are aimed at creating physiotherapy equipment based on the processing of ECG, EEG, glucometry, pulse waves, biologically active points and have prospects for independent application separately in the areas of treatment of heart diseases, brain diseases and the circulatory system;
• based on the results of the work, two candidate dissertations were defended, two methodological manuals with the seal of the Ministry of Education and Science of Ukraine were published, 10 scientific articles were published, 7 reports were made at international conferences;
• the results of the work are used in the educational process in six disciplines in the lecture process, laboratory and practical classes, when performing diploma theses of bachelors, specialists and masters, as well as in the work of postgraduate students;
• students of the specialization "Physical and Biomedical Electronics" were involved in the work, participated in the development of electronic layouts and software modules.

Goal and subject of work: development of non-invasive technologies for express diagnostics of diseases of the heart, brain, blood and blood vessels based on new methods and means of high-resolution measurement and data processing for the development of non-invasive technologies for automated diagnostics in medicine for faster and better provision of the doctor with the necessary information for early detection of pathologies, improvement of the treatment process, while reducing the number of invasive tests and procedures.

Performance
but scientific substantiation and implementation of methods for obtaining objective information with high resolution for assessing the functional state of a person, selection of a sufficient set of quantitative parameters for a well-founded diagnosis; mathematical models of systems that form informative signals of the physiological state of a person have been developed; a methodology for building high-resolution systems from the standpoint of reliability, accuracy, and reliability of their parameters has been developed. The main scientific idea: the use of new mathematical methods for processing diagnostically important signals based on high-resolution means in the time and amplitude spaces of the existence of signals. At the same time, using new models of physical phenomena in living organisms and the most advanced mathematical processing methods, diagnostic features are obtained that cannot be achieved using currently available technologies. To implement the use of high-resolution technologies in non-invasive glucose analysis, new methods of spectral analysis of glucose absorption during transcutaneous passage of infrared radiation have been applied. The synthesis of optimal filters was performed, a method for determining glucose concentration was developed. The principles of pattern recognition of late atrial potentials in the composition of electrocardiographic signals were developed, based on the selection of diagnostic features of late atrial potentials in the coordinate basis of the main eigenvectors of the covariance matrix of the electrocardiogram ensemble, which allows automating the process of classifying heart diseases by a rational number of features. An algorithm for spectral evaluation of low-amplitude components of electrocardiograms was created in the basis of eigenvectors of the noise subspace of the covariance matrix of the electrocardiogram ensemble, which provides higher spectral resolution than classical methods.

Based on the eigenvector method, methods for adaptive construction of standards for classes of epileptiform complexes were improved. The proposed standards consist of a generic complex of the class of epileptiform oscillations and the parameters of its possible permissible distortions, which allows for the temporal localization of epileptiform complexes in the electroencephalogram by means of pattern recognition using contextual information. An algorithm for adaptive construction of standards for classes of epileptiform oscillations in the electroencephalogram has been developed, which allows for automated pattern recognition of epileptiform complexes in the electroencephalogram by comparison with the standard and using the developed measure of proximity - significant maximum deviation.

To implement the use of high-resolution technologies in non-invasive glucose analysis, new methods of spectral analysis of glucose absorption during transcutaneous infrared radiation were applied. The synthesis of optimal filters was performed, and a method for determining glucose concentration was developed.

An ECG analysis method has been developed to detect late potentials of the SPP, which is a modification of the eigensubspace method. Application of the proposed method to improve the algorithmic support of the diagnostic system Application of the proposed method to improve the algorithmic support of the diagnostic system of the ECG BP based on the transformation into a coordinate basis of the eigenvectors of the covariance matrix of the ECG ensemble and the division of the full transformed space into signal and noise subspaces allows: to expand the dynamic range of detecting low-amplitude signals of late atrial potentials by increasing the signal/noise ratio; to select a rational number of diagnostic features of late atrial potentials by using decomposition coefficients in the basis of the main eigenvectors; to increase the resolution of the spectral evaluation of electrocardiographic signals for the purpose of detecting late atrial potentials by introducing functions that include eigenvectors of the noise subspace.

Based on the development of the eigensubspace method for a high-resolution ECG system, the following have been created: - a patient-adaptive filtering algorithm based on the restoration of electrocardiographic signals in the basis of the main eigenvectors, which allows to isolate useful low-amplitude late atrial potentials against the background of noise with minimal deformation of ECG waves and complexes; - combined algorithms for multi-channel compression and restoration of the ECG by sequentially applying one-dimensional orthogonal transformations in columns and rows, which make it possible to reduce the amount of data by 3-4 times when transmitting information; - algorithms for recognizing patterns of late atrial potentials in the ECG based on the selection of a rational number of diagnostic features in the transformed basis of the main eigenvectors.

A methodology for analyzing high-resolution electroencephalograms (EEG) using pattern recognition tools has been developed, which will allow to automate the early diagnosis of epilepsy. A new way of adaptively constructing a standard for class e is proposed.
of pileptiform complexes (EC) based on the use of the main vector of the matrix of averaged correlations for the matrix of the class of complexes. The standard of the developed type consists of a generic complex of oscillations and sets of parameters that allow recognizing ECs that have distortions in the real EEG signal, namely: stretching and compression of the complex, changes in its amplitude, displacement of the complexes relative to the isoelectric line, the presence of a trend and changes in the ratios between the amplitudes of the parts of the complex. The use of the standard constructed by this method allows obtaining standards for EC classes that are considered by the doctor to be similar to existing complexes, using contextual and a priori information and selecting the parameters of the standard.

A method for synthesizing an optimal filter for non-invasive glucose measurement is proposed. Based on the theoretical justification and calculation, the sensitivity function for the glucose measurement channel is obtained. It is shown that due to the low level of the glucose signal in the presence of other components, the success of obtaining reliable information for non-invasive glucose analysis is possible when using high resolution in signal amplitude and spectral selectivity. The performed analysis of the measuring channel showed the main requirements for the optical and measuring paths of the system for non-invasive determination of glucose content in human blood. The main problems of constructing the optical part of the system are shown, in particular, the lack of lasers of the required infrared range and the lack of width of the frequency tuning of the analysis. The justification for the choice of laser parameters is given. As an implementation of optical information capture for the instrumental implementation of the proposed glucose measurement method, an acousto-electric measurement method is proposed and patented, as well as the use of special ATR prisms. It is shown that increasing the resolution of the measuring path is fundamentally necessary for the practical implementation of non-invasive glucose measurement.

The theoretical foundations of high-resolution technology in non-invasive early and express diagnostics of cardiovascular system diseases, brain and non-invasive glucose analysis in diabetic patients are considered. The current state of the issue, the best world achievements and development prospects are shown. The sources of errors, methods for their reduction, software computational methods for signal processing that meet the requirements set by the use of high-resolution technology are considered. The results of the application of high-resolution methods in digital filtering, signal recognition and disease diagnostics are shown. The principles of building high-resolution technical means for non-invasive express diagnostics technologies are considered. The results of testing an experimental high-resolution system are presented. The modeling of medical signal identification in a high-resolution system is performed. The results obtained confirmed the high efficiency of using high-resolution systems in express diagnostics of early heart and brain diseases, in the treatment of patients with diabetes and acupuncture examinations.

The purpose of the work is to develop mathematical methods, algorithms and software for processing biomedical signals and images in automated complexes for early diagnosis of the cardiovascular and nervous systems in the treatment of the most common systemic human diseases. The purpose of the work is to increase the efficiency of computer systems for diagnosing the cardiovascular and nervous systems and to implement the developed methods and algorithms for processing biomedical signals and images in the production of diagnostic equipment. The research is aimed at creating and using the methodology and technical implementation of new mathematical methods for processing and analyzing medical signals and images in the presence of high levels of interference, noise and other artifacts. Complex mathematical processing of measurement information will allow for optimal diagnostics based on non-invasive technologies.

The relevance of the work lies in solving complex theoretical and practical issues of improving the algorithmic and software support of automated diagnostic systems for diseases of the cardiovascular and nervous systems. Existing signal processing methods: spectral, correlation, structural, convolutional, information, wavelet, etc., have their own characteristics both in use and in the results obtained, which often require additional analysis and evaluation.

In addition, the development of new medical equipment for early diagnosis of the cardiovascular and nervous systems requires a significant increase in the qualitative and quantitative indicators of medical information processing in
real time.

It was determined that the main problems of fetal electrocardiogram (ECG) analysis are the lack of necessary databases, low signal-to-noise ratio of fetal ECG, insufficient knowledge about the development and function of the fetal heart. Monitoring of the fetal condition is completely based on the fetal heartbeat and does not include the characteristics of the ECG signal waveforms, which are the cornerstone of assessing cardiac activity in children and adults. The main reason for the exclusion of this most important source of information from clinical practice is that the technology for reliable measurement of the ECG signal is largely unavailable. Most heart defects have some manifestations in the morphology of cardioelectric signals, which are recorded on the ECG and are believed to contain much more information compared to traditional ultrasound methods.

Technologies for separating maternal and fetal electrocardiograms based on independent component analysis, blind source separation based on models of electrocardiogram signals monitoring the electrical activity of the heart have been developed. Fetal cardiorhythmograms are a dependence that reflects both fast and slow changes in the fetal heart rate. These changes reflect components that characterize general and local manifestations of irregularity of the heart. In particular, there are possible manifestations of acceleration and deceleration of the rhythm according to arrhythmic, chaotic and extrasystolic variants.

For a numerical description of informative and potentially diagnostically useful changes in the characteristics of fetal heart rate variability, two approaches were proposed: using the decomposition of the signal of the dependence of heart rate on time into components, and determining the intervals of fetal rhythm stationarity based on statistical tests.

When determining stationarity, the concept of stationarity in a narrow sense was involved, and an assumption was made that signal areas can be considered as samples of some random process with an unknown probability density distribution law. The Wilcoxon rank criterion was used to determine the equality of median values.

A new method of visualizing the results of determining stationarity at different scales is proposed, which allows obtaining an integral idea of ​​the duration of stationarity intervals for the entire signal, as well as determining areas and regions in which the signal is stationary. An experimental software sample has been created to assess the parameters of fetal and maternal heart rate variability.

An assumption has been made regarding the presence of separate processes of different duration in the body that affect changes in heart rate variability. A method for decomposing the heart rate signal into components of different durations based on wavelet signal decomposition using standard wavelet functions and the possibility of using adapted maternal wavelets has been developed.

Purpose of the work: Creation of information support for modeling promising semiconductor nanometer devices for ultrafast integrated circuits in the gigahertz and terahertz ranges and information support for their manufacturing technologies.

Purpose of the work: Study of physical processes and physical-topological modeling of nanostructured electronic devices: multilayer heterotransistor structures, including heterobipolar transistors, heterotransistors with quantum dots, devices based on superlattices and structures with resonant tunneling for the development and implementation of their manufacturing technologies.

As a result of the work, theoretical foundations were created for the development of promising semiconductor nanometer devices for ultrafast integrated circuits in the gigahertz and terahertz ranges and information support for their manufacturing technologies.

An analytical method for calculating relaxation times for various types of carrier scattering was developed in semiconductor compounds AIIIBV, which allows for the study of their field-velocity and field-temperature characteristics. Based on this method, an analysis of dynamic parameters for the most promising semiconductors AIIIBV (GaAs, InP, GaN) in a strong electric field was carried out.
and, including high-frequency and dynamic conductivity and its limiting frequency. It was first established that the maximum frequency of existence of negative dynamic conductivity for gallium nitride is about 500 GHz.

Physical processes were studied and physical and topological modeling of nanostructured electronic devices was carried out: multilayer heterotransistor structures, including heterobipolar transistors, heterotransistors with quantum dots, devices and structures with resonant tunneling for the development and implementation of their manufacturing technologies. Mathematical models were developed for the analysis of relaxation parameters of 3D- and 2D-electron gas in semiconductor compounds AIIIBV in a strong electric field. Calculations of the main parameters for studying the dynamic properties of electron gas in multi-valley semiconductors were carried out, modeling of electron transport processes in nanometer heterostructures in strong electric fields was carried out and parameters of heterostructures for various purposes were optimized, promising structures of heterotransistors with multilayer channels, in particular, several heterojunctions and quantum dots, and resonant tunneling diodes were proposed.

A model of a resonant tunneling diode was created, which allows tracking the change in its characteristics depending on the design parameters - the thickness of the barrier layers and the layer from which the potential well is created; the molar fraction of Al in the AlGaAs compound from which potential barriers are formed; the temperature of the crystal lattice. For the developed model, scattering in DBCS was taken into account through the empirical pulse relaxation time. For the test structure, a study was conducted, as a result of which dependencies were obtained that allow predicting and selecting the necessary RTD topologies so as to obtain the desired I-V characteristic. Due to the effective ratio between the adequacy of the model and its mathematical complexity, sufficiently accurate results were obtained, which at the same time were relatively easy to interpret, which was done.

Mathematical models were developed for analytical calculations of current-voltage characteristics and small-signal parameters of heterotransistors based on the results of two-dimensional physical-topological modeling. The parameters of the structures and approximations obtained on the basis of two-dimensional modeling can be used in packages for circuit design of submicron EHF ICs, optimization of parameters of semiconductor structures and their technology.

Purpose of the work: The purpose of the work is to calculate the parameters of semiconductor materials based on theoretical initial data and construct dependencies based on the calculation results, as well as analyze the obtained data.

Purpose of the work: This work is devoted to the analysis of the parameters of the band structure of gallium nitride, characteristic scattering mechanisms and calculation of the corresponding relaxation times, which allows calculating the most important (field-velocity and field-temperature) electrophysical characteristics for three- and two-dimensional electron gas. These results were used to model physical processes and parameters of submicron and nanometer structures of resonant tunneling diodes, heterostructure transistors, and quantum dot transistors.

A study was conducted based on a system of fundamental equations for charge carrier transport, supplemented by a substantiated system of relaxation equations for conservation of momentum, energy, and particles for the analysis of unsteady drift, taking into account various types of scattering, which allow analyzing the speed properties of transistor structures.

A two-dimensional modeling technique for transistors with high electron mobility was developed and improved, which allows describing submicron effects: drift velocity "spike", heating in a strong electric field, intervalley transfer, quasi-ballistic effect.

The purpose of the work is to create information support for modeling promising semiconductor nanometer devices for ultrafast integrated circuits of the gigahertz and terahertz ranges. Mathematical models have been developed, and analysis of resonant tunneling diode and nanotransistor structures is aimed at
search for ways to increase performance based on energy, field, and other dependences of charge carriers for the use of these structures in submicron and nanometer circuits.

The result of the work is the calculated parameters and characteristics of trinitrides, describing their dynamic properties in a strong electric field, and programs for modeling the studied heterostructural nanocomponents have been developed.

The result of the work is also the possibility of using mathematical models for the analysis and further optimization of the above-mentioned active nanocomponents, based on the properties of materials, topology of structures, parameters of heterojunctions, tunnel barriers, characteristics of nanoscale regions of structures, doping level, etc.

The purpose of the work is to create information support for modeling promising semiconductor nanometer devices for ultrafast integrated circuits in the gigahertz and terahertz ranges and information support for their manufacturing technologies.

The subject of the work is the study of physical processes and physical-topological modeling of nanostructured electronic devices: multilayer heterotransistor structures, including heterobipolar transistors, heterotransistors with quantum dots, devices and structures with resonant tunneling for the development and implementation of their manufacturing technologies.

The work is aimed at mathematical modeling to create theoretical foundations and develop promising semiconductor nanometer devices for ultrafast integrated circuits in the gigahertz and terahertz ranges and information support for their manufacturing technologies.

The proposed methodology and software tools for modeling the characteristics of a wide range of semiconductor nanometer devices; including heterobipolar transistors, heterotransistors with quantum dots, devices and structures with resonant tunneling: topology and designs of promising semiconductor nanometer devices.

The mathematical models, algorithms and programs proposed in the work contribute to the creation of modern semiconductor components for telecommunication systems and signal processing systems, the development of nanotechnologies.

Aim and purpose of the work: This work is devoted to the analysis of the parameters of the band structure of gallium nitride, characteristic scattering mechanisms and calculation of the corresponding relaxation times, which allows calculating the most important (field-velocity and field-temperature) electrophysical characteristics for three- and two-dimensional electron gas. These results will be used to model physical processes and parameters of submicron and nanometer structures of resonant tunneling diodes, heterostructure transistors, and quantum dot transistors.

The work calculates the parameters and characteristics using the method of averaging the rates of various types of scattering, which has so far shown good agreement with experimental data.

The work investigates the characteristics and parameters of the band energy structure of gallium nitride and the influence of the parameters of this structure on the electrical properties of the material.

Taking into account the importance and peculiarities of carrier transport effects in a solid, the mechanisms of collisions leading to the scattering of directed carrier flows have been investigated. It has been proven that electron-electron and electron-hole scattering are prominent in the excitation defects in the energy spectrum of electrons (plasmons, excitons), however, the most typical are impurity scattering and various types of phonon scattering. By averaging the scattering rates over the distribution function, the average relaxation times of momentum and energy are found.

The report also considers the transport properties of gallium nitride - the conditions that determine the mobility of carriers. The presented results of modeling the mobility of charge carriers and analyzing the contribution of each type of scattering to the mobility value prove that one of the main features of scattering in gallium nitride is associated with the contribution of the ionic bond between Ga and N. This leads to the fact that among the phonon types of scattering, the polar optical one plays the main role. The second is
The peculiarity is associated with the presence of intervalley transitions. Gallium nitride has a single lower G-valley and six higher X-valleys, and eight higher L-valleys. The presence of an intervalley transition is detected at higher temperatures. At temperatures above 600 K, transitions from G-valley states to X-valley states are possible. Acoustic scattering in gallium nitride is quite weak and is detected in the absence of competing types of scattering: impurity and polar optical. Its small manifestation can be seen at temperatures above 600-700 K.