One of the main trends in microwave electronics is the ultra-large power production. The electron stream energy is converted inside vacuum systems, where the key moment is increasing output power of microwave devices, which is possible only when using more and more powerful electron streams. Increasing electron stream power is possible due to either enhancing the carried currents or as a result of increasing the electron energy. Given the law that connects currents and voltages in electronic systems operating when the current is limited by a spatial charge, the production of ultra-high-power electron flows is associated with the usage of relativistic velocity electrons, i. e. approaching the light speed. Likewise, at present, relativistic electrovacuum devices (traveling-wave lamps and backward-wave lamps) use magnetic focusing for linear relativistic streams, which prevents the implementation of simple superconducting electrodynamic systems, because highfrequency metal superconductivity disappears in constant magnetic fields. Meanwhile, simplified ultra-highpower superconducting device structures can significantly increase the device energy due to the strong ohmic loss reduction, which just limits the device energy, destroying the working electrodynamic system surface by increasing power or pulse duration of the generator. The article outlines the modernized design of a new-type microwave generator – the relativistic helitron. The paper considers a simpler coaxial resonator design, obtained by using the supercritical narrowing of the inner conductor radius by the H_n1l mode of the electromagnetic field, rather than a coaxial resonator with notch filters.

In modern microwave electronics, it is extremely important to evaluate the possibility of efficient generation of highly stable oscillation modes and the creation of low-noise generators with a millimeter wavelength range. The purpose of the research was to determine the type of wave for the most effective modes of oscillation and to ensure the operation of various modes in the mode of mutual synchronization, which becomes possible because the real design of generators can include a number of reactive elements that form a number of resonant frequencies in the microwave range. This paper shows how it is possible to provide conditions for generation at a particular frequency and simultaneously obtain high stability of vibrations.
The main research methods were: analysis and generalization of literature data and results of our own research on the peculiarities of oscillation generation in a multi-circuit system of a Gann diode generator and conditions under which high frequency stability is established; experimental studies using microwave spectrum analyzers, microwave devices. In this paper, the possible generation frequencies are calculated, the operation modes of the diode generator are studied, and their phase noise is experimentally measured. It is shown that when the generation frequency approaches the cut-off frequency of the waveguide with the help of structural elements or by changing the supply voltage, a mode of mutual synchronization of vibrations at high harmonics occurs using the H₁₀₂ wave type in a very narrow range of frequency (phase) noise. A generator with a multicircuit resonant system was experimentally studied and a small amount of phase noise was obtained for both types of H₁₀₁ and H₁₀₂ waves. At the same time, the best parameters and noise characteristics occurred at the lowest types of vibration types.
These results have an experimental novelty, which allows us to hope for the development of constructive solutions that provide the creation of highly stable low-noise generators of millimeter wavelength range with the minimum mass dimensional parameters.

Structures based on ZnO and Cu, which are a polycrystalline composite consisting of crystalline ZnO with a crystallographic orientation of (002) and (101) doped with Cu and crystalline metallic Cu, were obtained by electrochemical deposition on substrates of single crystal origin. In the study of the obtained films by Raman spectroscopy, the forming of crystalline ZnO was confirmed. ZnO 2A₁ (LO), also in the spectra of each of the bands present in the 649 cm^-1 region, not related to the vibrating lattice modes of the intrinsic crystalline ZnO. It was shown that with an increase in the deposition current density in the range of 2–10 mA/cm², the concentration of Cu in the material weakens, while the number of Cu clusters decreases, but the degree of doping of ZnO with Cu ions increases. ZnO-based composites exhibit a broad photoluminescence band in the long wavelength range of 500–700 nm, related with vacancies and interstitial oxygen atoms in the crystal lattice. At a current density of 5 mA / cm², short-wavelength shifts of the photoluminescence bands are observed, due to the doping of Cu, since impurity levels are created in the band gap associated with the presence of Cu in ZnO films. A change in the radiation intensities was observed at a current density of 10 mA/cm², which is due to the greater thickness of the obtained films. The results can be used to develop the manufacturing technology of optoelectronic and photovoltaic devices, photocatalytic coatings based on ZnO.

The purpose of this paper is to develop a method to enhance reliability of vibrational diagnosing of variable-speed industrial equipment. A dataset of vibration acceleration signals, picked up at the variablespeed test stand, has been obtained. Preliminary splitting of a vibration signal into three frequency ranges has been proved to be necessary. Vibration power dependencies on the main shaft speed have been estimated in different frequency ranges. The paper proposes an algorithm compensating variations of instantaneous power of vibration signal where equipment operation speed varies. It is based on the use of empirical dependencies of vibration signal power on shaft speed, which were derived by ensemble averaging of preliminarily split signals. A root mean square (RMS) value calculated in a sliding window is used to estimate variation of signal power. For signals within each frequency range, ensemble-averaged relative power variation produced by speed deviation is to be estimated. Instantaneous values of signals in each frequency range are to be divided by relations estimated as above. Thus, only power variations caused by variable speed are compensated. Variations caused by defect evolution are preserved. The resulting signal to be further analysed is derived by summation of processed signals in three frequency ranges. Power variation compensation decreases dispersion of parameters of signal that are used for estimation of equipment state. Preliminary compensation of vibration power variation caused by variable operation speed has proved to be effective for improving vibrational diagnostic system results. The proposed algorithm was validated on such statistical parameters of vibration as RMS and peak-factor of vibration signal.

This article studies and analyzes the results of applying numerical iterative methods for solving nonlinear equation systems (Newton, modified Newton's method, gradient descent, sequential iterations, Levenberg – Marquardt), compiled and used to calculate the rectangular spatial coordinates of radio emission sources in range-difference passive radars of various configurations (incorporating from 3 to 4 receiving points). The aim of the research was to determine the optimal number of receiving points and to select the most effective algorithm for coordinate transformations of the vector of observed parameters (a set of range difference estimates from radio emission sources to the corresponding pairs of receiving points) into the vector of measured parameters (rectangular spatial coordinates). The following parameters were used as comparison criteria: passive radar working area (a part of space where the deviation of target coordinate estimates from their true values does not exceed the maximum tolerable values); average error in calculating spatial coordinates in the working area; iterations number of coordinate calculation in the analyzed part of space. Upon completing a comparative analysis of obtained characteristics and dependencies, we concluded that it is optimal to include four receiving points in a range-difference passive radar and use the Levenberg – Marquardt method to calculate the spatial coordinates of radio emission sources.

This article proposes an approach for designing a general subsystem of automatic generation of questions in intelligent learning systems. The designed subsystem allows various types of questions to be automatically generated based on information from the knowledge bases and save the generated questions in the subsystem knowledge base for future use. The main part of the subsystem is the automatic generation module of questions, which allows one to generate questions of various types based on existing question generation strategies in combination with the structural characteristics of knowledge bases built using OSTIS technology. In this article, a variety of strategies for automatically generated questions are proposed, the use of which allows various types of questions to be automatically generated, such as multiple-choice questions, fill-in-the-blank questions, questions of definition interpretation and etc. The most important part of the subsystem is the knowledge base, which stores the ontology of questions, including the question instances themselves. In this article, the knowledge base is constructed based on OSTIS technical standards. The type classification of automatically generated questions was developed, as well as the subject area for storing generated questions and the corresponding ontology described in the knowledge base of the subsystem. The generated questions are stored in the subsystem knowledge base in the form of SC-code, which is the OSTIS technology standard. When testing users, these automatically generated questions are converted to the corresponding natural language form through the natural language interface. Compared with the existing approaches, the approach proposed in this article has certain advantages, and the subsystem designed using this approach can be used in various OSTISbased systems driven by OSTIS technology.

The purpose of the work is to develop models and algorithms for optimizing matching in dynamically generated graphs of asymmetric relations in coordinated open systems of interacting agents with centralized and collective control. The dynamic asymmetric matching optimization problem arises here as a result of a compromise approximation of the mapping of the dynamic programming method onto a stream of known open assignment problems or several traveling salesmen. However, the branching alternatives presented in this way for independent tasks do not take into account the interdependence of real relationships between agents and their tasks, including their relationship to time. Ignoring the dependence of branching alternatives leads to a delay in the moment or to a loss in the quality of assignment of tasks to coordinated agents. The main idea of the proposed implementation of the principle known for effective control is to postpone the moment the final decision is made to the latest moment, taking into account the susceptibility of the system to local changes in state variables. The interdependence of states is revealed on the basis of the analysis of the correspondence of the graph of the current matching with the optimal solution on the subgraph of perfect matching. The transition between states is implemented by the incremental version of the reoptimization algorithm for solving linear problems of assigning the shortest replenishing path using the method. The space of search states is a dynamically generated bipartite sparse graph of alternatives for a combination of agents and tasks, represented by a list of arcs. To highlight the sets of changed arcs, it is proposed to supplement the weight of the arcs with the boundaries of the stability intervals of the solution, optionally formed in the background. By default, the weight of the modified arc matches the boundary of the stability interval. On each correction cycle of the lists of agents, tasks, and their associations, subsets of elements are selected for which reconsideration of matching is required. An enhanced condition for the selection of such elements is to go beyond the boundaries of the stability interval. In this case, the asymmetry of the assignment problem is taken into account by choosing the adjacency structure for the fraction of the graph with a minimum of vertices. As a result, the reaction time of procedures for solving the assignment problem is reduced by an order of magnitude.

The aim of the study was to develop a methodology for assessing the state of motor adaptation at the level of the main joint elements of the locomotor system when performing postural deviations with inertial components in a group of healthy volunteers (n=24). To conduct the study we used the “Teslasuit” smart suit as a technology with a system of inertial measuring units. A virtual skeletal model of the subject’s body was reconstructed on the obtained quaternions for each direction of spatial displacement. Parameters of inertial kinematic were calculated by the Fast Fourier Transform in the frequency bands of 0.1–5, 6–10, and 11–15 Hz. To assess motor adaptive reactions, we developed the following tests: ventrodorsal displacement test; laterolateral displacement test; linear displacement test in vertical direction; axial rotation test around vertical. All test tasks were performed using biofeedback as a virtual reality environment. The study revealed the presence of universal motor adaptation mechanisms with activation of the components of axial rotation of the trunk and axial rotation and flexion of the leading shoulder joint. At the same time, a dynamic phase of postural regulation during axial rotations and tilts of the body leads to the activation of motor adaptation mechanisms from the leading hip, knee, and ankle joints, while axial movements form a picture of the kinematic stabilization of these locomotor system elements.

The purpose of the work is the development of basic data structures, speed-efficient and memoryefficient algorithms for tracking changes in predefined decisions about sets of shortest paths on transport networks, notifications about which are received by autonomous coordinated transport agents with centralized or collective control. A characteristic feature of transport operations is the independence and asynchrony of the emergence of perturbations of optimal solutions, as well as the lack of global influence of individual perturbations on the set of all processes on the network. This clearly determines the feasibility of realizing the idea of reoptimizing existing solutions in real time as information is received about disturbances in the structure and parameters of the transport network, various restrictions on the use of existing shortest paths. In contrast to the classical problems of finding shortest paths on static or dynamic graphs, it is proposed to supplement the set of situations controlled by the observer by taking into account the associations of shortest path trees with agents that actually use such paths. This will improve the responsiveness of agent notification processes for timely switching to a new path. The space of search states is a dynamically generated bipartite sparse graph of the transport network, represented by a list of arcs. The basic algorithm for finding the shortest paths uses Dijkstra's scheme, but implements a bootstrapping method to generate the search result. The compactness of the representation of the observed forest of shortest paths is achieved by mapping individual trees of such a forest onto the projection of tree vertices in memory, where the position of each vertex corresponds to the distance from the tree root. The proposed version of the construction of the search procedure is based on the mechanisms existing in database management systems for creating different relational representations of the physical data model. This eliminates the need to solve technological problems of complexing heterogeneous models of dynamic transport networks, memory allocation. As a result, the specification of various rules for the logistics of transport operations is simplified, since such operations in terms of object-oriented models are easily determined by polymorphic classes of transitions between nodes of the transport network.

Individual forecasting of the reliability of semiconductor devices, taking into account gradual failures, is an urgent task, as it allows you to choose highly reliable instances for critical electronic devices of long-term functioning. In relation to bipolar transistors, an approach is proposed that allows us to solve this problem by using the voltage applied to the collector-emitter junction as a simulated effect. Using the example of highpower bipolar transistors of the KT872A type, it is shown how the problem is solved. For the sample of transistors of this type using the results of a training experiment, two equations were obtained to describe the electrical parameter under consideration (a static base current transfer coefficient in a circuit with a common emitter), the value of which judges the absence or presence of a gradual failure for a specific instance. The first equation shows how the electrical parameter changes on average depending on the voltage applied to the collector – emitter junction. The second equation describes the average degradation of the electrical parameter during long-term operating time of transistors. Based on these two equations, a simulation model of the reliability of bipolar transistors of the type in question is obtained in the form of a communication function that shows what level of simulation voltage corresponds to a given operating time. As applied to the transistors of the type under consideration, the obtained simulation model allows us to individually predict reliability by the gradual failures of the same type of samples that did not participate in the training experiment. To do this, first determine the value of the simulation voltage corresponding to a given operating time. This is achieved by substituting a given operating time into the model. The individual forecasting of the reliability of a new onetype instance consists in measuring the electrical parameter of this instance at a voltage on the transistor collector corresponding to the calculated simulation value, and comparing the measurement result with the norm set on the electrical parameter.

The purpose of the work outlined in the article is to review and demonstrate the use of graph technologies for deep data analysis. The first part of the article discusses the Intelligent System for the Comprehensive Analysis of Internet Sources Data and its possible directions for its further development. This system is a multi-purpose cluster using technologies for constructing a knowledge graph, methods and models of machine learning for in-depth analysis of data from Internet sources (for example, scientific publications, social networks, media). The purpose of the analysis is to identify the most important publications in a certain area (for example, in robotics, space research, healthcare, in the social sphere), thematic analysis of these publications, to identify the leader of a scientific direction and to predict trends in the development of directions and interaction of groups of people. When developing this system, we utilized probabilistic machine learning algorithms and methods for constructing and maintaining a graph model of the social network of authors and their publications, determining the rating of a particular author, determining the topics of publications and classifying them by areas of knowledge. The basis for the creation of intelligent applications is graph technology, which allows you to make predictions that are more accurate. The combined application of methods and algorithms of machine learning with graph technologies allows you to get hidden dependencies and perform predictive analysis of information, get answers in real time, and implement artificial intelligence algorithms. Methods of collaboration with graph technologies and a learning machine (for example, using neural networks) are based on graph embedding. This technology allows you to perform a comprehensive, deep and intelligent analysis of information. At the end of the article, there are analytical reports obtained using graph technologies in the Intelligent System for Complex Analysis of Internet Sources Data.