The relevance of the study is the need to improve the reliability of non-contact control of the geometric parameters of fuel assemblies under conditions of limited access to the objects and exposure to ionizing radiation. The implementation of the parallax-shift method, which uses sequential frames acquired by a television camera during its arc-shaped motion to form quasi-stereo pairs of images, requires the selection of an imaging device that simultaneously provides sufficient spatial resolution, radiation resistance, temperature stability, and stability of image parameters. The aim of this study is to substantiate the selection of a charge-coupled device (CCD)-based television camera for a non-contact system for measuring the height differences of fuel assemblies under radiation exposure. Methods. Requirements for the main technical characteristics of the camera were established based on the operating conditions in the core of a water-cooled, water-moderated power reactor. A weighted-criteria method was applied to comparatively evaluate the alternatives, taking into account radiation resistance, operating temperature range, spatial resolution, geometric stability, and characteristics of the generated image. In addition, the influence of camera parameters on the accuracy of subsequent height-difference determination using the parallax-shift method was evaluated. Results. A comparative analysis of five television camera models was performed. The Hitachi KP-DE500R camera was identified as the most preferable option, with an integrated score of Q = 0,91. It was demonstrated that the selected camera provides the required television image quality and possesses the characteristics necessary for forming quasi-stereo pairs during arc-shaped camera motion. Verification of the measurement method using subpixel localization of informative image elements and the accumulation of approximately 250 quasi-stereo pairs demonstrated the feasibility of determining height differences with an accuracy of 1.3--2.0 mm. The novelty of the study lies in substantiating the selection of the type and specific model of a television camera for a measurement system based on the parallax-shift method, taking into account a combination of operational and metrological requirements. Practical significance. The results obtained can be used in the development of non-contact television inspection systems for monitoring the geometric parameters of fuel assemblies at nuclear power plants.

Problem statement (relevance). The digital educational environment of a specialized university, which has evolved as a set of loosely coupled automated systems, successfully handles administrative support and content delivery, but proves to be fundamentally limited in providing end-to-end competency analytics, personalized career guidance, adaptability to labour market dynamics, and diagnostics of its own integration integrity. Overcoming these limitations requires the introduction of a specialized educational technology -- IT‑Skills -- whose success depends on the validity of a system of information-functional principles. Earlier, the author proposed a two‑level system comprising five basic (necessary) and four optional (sufficient) principles. The validity of the latter requires rigorous proof. The aim of this work is to prove the validity of the principles of adaptability, personalization, analytics, and diagnosability for integrating IT‑Skills technology into the digital educational environment. To prove this, a modified reasoning-from-the-alternative method is applied: for each principle, an antithesis is constructed, and a cause-and-effect analysis reveals the degradation in the indicators of skill model currency, goal alignment accuracy, reporting completeness, and integration failure detection time. The novelty lies in the fact that for the first time the four optional principles are substantiated as sufficient conditions that transform integration from a mere technical act into a meaningful renewal of the digital educational environment. The result of the work is a proof of the necessity of adhering to the optional principles; a holistic regulatory framework for integration is established, which is hypothesized to meet the information-functional needs of all participants in the educational process. The theoretical significance lies in the formalization of integration effectiveness indicators through mathematical expressions; quantitative criteria are also introduced for verifying the effectiveness of each principle. The practical significance is that compliance with the substantiated principles collectively ensures an improvement in the quality of IT personnel training.

Telepresence is a key service in sixth-generation networks, placing stringent quality of service requirements on them. Current and future communications infrastructures are characterized by a high degree of heterogeneity. Ensuring a stable level of quality of service is therefore a complex task requiring an integrated approach and innovative solutions. Traditional management methods, such as static prioritization or adaptive algorithms, lack sufficient flexibility, while centralized machine learning approaches are associated with privacy risks and high network load. The aim of this paper is to systematize and comparatively analyze federated learning methods used to predict and optimize quality of service in heterogeneous telepresence networks. The essence of the approach proposed in this paper lies in the author's classification of federated learning methods based on two criteria: architectural organization and the method of adaptation to heterogeneity. Detailed descriptions are provided for each class, and strengths and weaknesses are highlighted. The comparative analysis is based on comparing approaches based on key metrics: forecast accuracy, convergence rate, communication delays, privacy, device computational load, heterogeneity resilience, implementation complexity, and security. The analysis was based on data from peer-reviewed sources, which allowed us to identify quantitative advantages (e.g., a 1.91--38.89% reduction in training rounds and a 9.52--40.00% reduction in communication delays for the hierarchical MultiFed architecture). Experimental data presented in the analyzed papers confirm that federated learning provides a balance between predictive accuracy, privacy, and scalability, with the gains being most noticeable for devices with heterogeneous data. The scientific novelty lies in the proposed classification of federated learning methods and the systematization of their advantages and limitations as applied to the specifics of telepresence networks, which enables an informed choice of architecture depending on the use case. The authors' practical significance lies in their recommendations for selecting a federated learning method for various conditions (regional heterogeneity, limited device resources, and stringent latency requirements), which can be used in the design and operation of communication networks.

Purpose. Modern and advanced wireless communication systems impose high requirements on the accuracy of device positioning, which makes the task of angle-of-arrival estimation relevant. The aim of the work is experimental approbation of angle-of-arrival estimation algorithms on the affordable software-defined radio platform KrakenSDR in a real urban environment. Methods. The paper considers five algorithms: the classical Bartlett spectral algorithm, super-resolution algorithms MVDR and MEM, as well as subspace algorithms MUSIC and Root-MUSIC. The experiment was conducted for two antenna array configurations (circular and linear) with subsequent post-processing in MATLAB. Elements of novelty include comprehensive experimental comparison of five angle-of-arrival estimation algorithms of different nature on a single low-cost SDR platform, as well as evaluation of the influence of antenna array configuration and processing mode (real-time and post-processing) on the direction-finding accuracy in urban conditions. Results. It was found that the MUSIC algorithm provides the most stable results, and the linear antenna array demonstrates lower error and scatter of estimates compared to the circular one. Post-processing of recorded data allows reducing direction-finding errors for most algorithms. For the circular array under favorable conditions, the mean absolute bias was 2.8–6.0°, and the standard deviation was 0.0–3.5°. Practical relevance. The presented results confirm the applicability of the KrakenSDR platform as an affordable experimental testbed for research on angle-of-arrival estimation algorithms. The obtained data can be used in the development of positioning systems, radio monitoring, and electronic intelligence based on low-cost SDR solutions.

Purpose. The increasing number of connected devices and the intensity of information exchange in Internet of Things systems leads to an increased load on communication channels and computing nodes. Centralized transmission of the entire data volume to the cloud infrastructure with limited channel bandwidth and finite buffer capacity may be accompanied by an increase in delivery latency and the probability of message loss. Distributing processing between edge, fog, and cloud nodes allows for a reduction in the volume of traffic transmitted to the upper network layers; however, the effectiveness of various architecture options depends on the intensity, structure, and variability of the input flow. The aim of this work is to evaluate the impact of distributing the processing of heterogeneous IoT traffic between the edge, fog, and cloud layers on the average delivery latency and the probability of message loss. Methods. The solution to the problem is based on discrete-event simulation in the AnyLogic environment using elements of the process modeling library. Three system design options are investigated: centralized cloud processing, joint use of edge and cloud computing, and a three-tier Edge-Fog-Cloud architecture. The model takes into account three message classes, finite buffer capacity, performance of computing nodes and communication channels, aggregation of a portion of the processed traffic, as well as the variability of message arrival and servicing processes. Novelty. The novelty of the presented study is the simulation implementation of a multi-tier architecture for processing heterogeneous IoT traffic with separate consideration of the edge, fog, and cloud layers, as well as a study of the impact of variability of arrival and servicing processes on message delivery delay with changes in the input flow intensity. Results. A simulation model has been developed that enables a parametric study of three architecture options with changes in the input flow intensity from 20 to 400 messages per second. The dependences of the probability of message loss and the average delivery delay on the traffic intensity for three processing scenarios are obtained. For a multi-tier architecture, dependencies of average latency on the degree of variability of the arrival and service processes were additionally obtained. Practical relevance. The developed model enables a comparative evaluation of distributed IoT system design options, identification of congestion zones, and assessment of the impact of traffic characteristics, computing nodes, communication channels, and final buffers on message service quality.