Artem Kazarian

This paper addresses the critical problem of user identification within the framework of human-computer interaction (HCI). A novel identification method is proposed based on users' "polyfactor portraits." These portraits effectively reflect how individuals subjectively perceive specific HCI objects. The primary objective of this approach is to model the irrational components of human behavior, thereby improving system automation and intellectualization. Structurally, the developed method integrates a specialized identification algorithm with a core supporting mathematical model. To validate the efficiency of the proposed approach, a comprehensive practical experiment was conducted. This experiment involved the identification of technical support team members based on the dominant features of their interaction with a specific supported software product. The experimental results confirm the applicability of the method. Finally, the paper outlines future research directions aimed at potential improvements, technical enhancements, and broader practical applications of this identification method in real-world systems.

The article presents the architecture, model, and functional capabilities of a system for automated generation, translation, and publication of content based on large language models. The relevance of the development is determined by the needs of businesses, educational platforms, and socio-political projects for the rapid creation of multilingual materials, as well as by the necessity to counteract disinformation campaigns. The proposed solution is distinguished by its versatility and ease of deployment, even in resource-constrained environments, surpassing existing analogues that typically cover only individual stages of content management. Future development of the system envisions the integration of monitoring, automatic scaling, and the implementation of multi-level role management.

Driving safety education remains a critical societal priority, and understanding traffic rules is essential for reducing road accidents and improving driver awareness. This study presents the development and evaluation of a virtual simulator for learning traffic rules, incorporating spherical video technology and interactive training scenarios. The primary objective was to enhance the accessibility and effectiveness of traffic rule education by utilizing modern virtual reality approaches without the need for specialized equipment. A key research component is using Petri net-based models to study the simulator’s dynamic states, enabling the analysis and optimization of system behavior. The developed simulator employs large language models for the automated generation of educational content and test questions, supporting personalized learning experiences. Additionally, a model for determining the camera rotation angle was proposed, ensuring a realistic and immersive presentation of training scenarios within the simulator. The system’s cloud-based, modular software architecture and cross-platform algorithms ensure flexibility, scalability, and compatibility across devices. The simulator allows users to practice traffic rules in realistic road environments with the aid of spherical videos and receive immediate feedback through contextual prompts. The developed system stands out from existing traffic rule learning platforms by combining spherical video technology, large language model-based content generation, and cloud architecture to create a more interactive, adaptive, and realistic learning experience. The experimental results confirm the simulator’s high efficiency in improving users’ knowledge of traffic rules and practical decision-making skills.

This article presents the development and implementation of a web-based intelligent assistant for generating optimal clothing combinations based on real-time weather conditions and individual style preferences. The core innovation lies in the integration of large language models (LLMs), specifically OpenAI’s GPT-based API, which enables the automated classification, description, and compatibility assessment of clothing items. The system operates as a personalized wardrobe assistant that leverages cloud-based data storage, modern front-end technologies, weather APIs, and semantic analysis to deliver stylistically coherent and weather-appropriate outfit recommendations. The proposed solution addresses the growing need for intelligent wardrobe management and outfit planning tools by automating decision-making processes that traditionally require personal judgment or domain-specific expertise. Users can upload images of their clothing items, which are processed using the LLM to extract relevant attributes including type, subtype, style, color, weather suitability, and gender. The processed data is stored in Firebase Firestore and serves as the basis for generating personalized outfit recommendations. A key component of the recommendation algorithm is the compatibility matrix, which evaluates the stylistic and color harmony between various clothing elements. Compatibility scores are pre-defined based on common fashion principles and empirical observations. The algorithm calculates a weighted score for each possible outfit by considering both structural (e.g., shirt with jeans) and visual (e.g., black with white) compatibility. It also incorporates current weather data obtained via the Open-Meteo API, which includes parameters such as temperature, wind speed, and precipitation. These values are transformed into descriptive weather categories (e.g., cold, hot, windy) to refine the selection of suitable clothing items. The system’s user interface is built with HTML, CSS, Bootstrap 5, and JavaScript, providing a responsive and intuitive environment. Users interact with the system through several dedicated pages: uploading and classifying clothes, browsing the wardrobe, viewing local weather forecasts, and generating optimal outfit suggestions. The architecture supports drag-and-drop functionality for image uploads, automatic geolocation, real-time API interactions, and dynamic filtering based on weather and style input. The recommendation algorithm performs multi-stage processing: initial weather and style filtering, grouping clothes by type (top, bottom, footwear, outerwear), calculating all feasible combinations, and scoring them using compatibility formulas. The five best-scoring combinations are presented to the user. This approach ensures that the generated outfits are not only weather-compliant but also visually and stylistically balanced. Testing of the web application demonstrated high accuracy in clothing classification and relevance of generated outfits. The results validate the feasibility of using LLMs for semantic garment analysis and highlight the potential of such systems in fashion consulting, virtual wardrobes, and e-commerce personalization. Furthermore, the modular design and cloud-based architecture facilitate future expansion, including social features and brand-based recommendations.

In the era of rapid technological advancement, when robotics and intelligent systems are becoming an integral part of everyday life, the importance of developing control systems for mobile robotic platforms using artificial neural networks becomes extremely high and relevant. This field not only has significant practical needs but also holds considerable potential for innovative development. The evolution of modern robotics and computational intelligence has necessitated the creation of more efficient and adaptive mobile robotic systems. A system and tools for controlling mobile robotic platforms using artificial neural networks (ANNs) have been developed in this work. By simulating the workings of a neural system, ANNs enable robots not only to react to input data but also to learn to solve complex tasks and adapt to changes in their environment. One of the key challenges in mobile platform control is the development of effective and intuitive interfaces that provide convenient and reliable interaction between the user and the robotic system. In this context, the use of hand gestures by humans represents a progressive and promising direction as it allows for the creation of the most natural and efficient means of control. The main task is to create an effective and intuitively understandable system that enables the operator to interact with the robotic platform using natural movements and gestures. As a result, software with a graphical interface for real-time gesture recognition using machine learning has been developed. The scientific novelty of the approach is the integration of advanced ANNs methods to improve the quality of control and functionality of mobile robotic platforms. The main aspects of scientific novelty include integration with artificial intelligence, interactivity of control, development of robotics mobility, and adaptability to various tasks. The problem addressed in this work lies in the need to develop effective and intuitive control systems for mobile robotic platforms using gesture recognition technologies. A program based on convolutional neural networks has been developed, which determines the position of the hand and identifies specific gestures such as forward, backward, right and left turns, as well as stopping movement. The implemented technology can be used in various fields of human activity (smart home control, technological solutions for people with physical disabilities, enhancing interactivity in entertainment devices, improving interfaces for interacting with technical equipment).

The article examines the efficiency of spherical video usage for road traffic rule learning and driving skill acquisition in comparison to conventional driving instruction methods. Additionally, it outlines the creation of a virtual educational simulator designed for road traffic rule comprehension and enhancement of driving abilities. The outcomes of this investigation hold significance for road safety and the design of driver education programs. A novel teaching model employing spherical video streaming for road traffic rule understanding has been devised. The research was provided in the scope of Erasmus+ Jean Monnet «Augmented Reality for Education: implementation of European experience» project.

A method for optimal structure selection of hidden layers of the artificial neural network (ANN) is proposed. Its main idea is the practical application of several internal structures of ANN and further calculation of the error of each hidden layer structure using identical data sets for ANN training. The method is based on the alternate comparison of the expected result values and the actual results of the feedforward artificial neural networks with a different number of inner layers and a different number of neurons on each layer. The method afforces searching the optimal internal structure of ANN for usage in the development of "smart" house systems and for calculation of the optimal energy consumption level in accordance with current conditions, such as room temperature, presence of people, and time of the day. The usage of the presented method allows to reduce the time spent on choosing the effective structure of the artificial neural network at the initial stages of research and to pay more attention to the relationship between the input and output data, as well as to such important parameters of the ANN learning process, as a number of training iterations, minimal training error, etc. The software has been developed that allows to carry out the processes of training, testing, and obtaining the output results of the algorithm of the artificial neural network, such as the expected value of power consumption and operating time of each individual appliance. The disadvantage of the approach used in finding the optimal internal structure of the artificial neural network is that each subsequent structure is created on the basis of the most efficient of the previously created structures without analyzing other structures that showed worse results with fewer hidden layers. It was found that to improve the solution of this problem it is necessary to create a mechanism which will be based on the analysis of input data, output data, will analyze the internal relationships between parameters and will optimize the network structure at each stage using certain logical rules according to the results obtained in the previous step. It is established that this problem is a nonlinear programming problem that can be solved in the further development of this study.

In the process of the “smart” house systems work, there is a need to process fuzzy input data. The models based on the artificial neural networks are used to process fuzzy input data from the sensors. However, each artificial neural network has a certain advantage and, with a different accuracy, allows one to process different types of data and generate control signals. To solve this problem, a method of choosing the optimal type of artificial neural network has been proposed. It is based on solving an optimization problem, where the optimization criterion is an error of a certain type of artificial neural network determined to control the corresponding subsystem of a “smart” house. In the process of learning different types of artificial neural networks, the same historical input data are used. The research presents the dependencies between the types of neural networks, the number of inner layers of the artificial neural network, the number of neurons on each inner layer, the error of the settings parameters calculation of the relative expected results.

An appearance of radiometers and dosimeters on free sale made it possible to provide better radiation safety for citizens. The effects of radiation may not appear all at once. They can manifest themselves in decades to come in future generations, in the form of cancer, genetic mutations, etc. For this reason, we have developed in this paper a microcontroller-based radiation monitoring system. The system determines an accumulated radiation dose for a certain period, as well as gives alarm signals when the rate of the equivalent dose exceeds. The high reliability of this system is ensured by a rapid response to emergency situations: excess of the allowable power of the equivalent radiation dose and the accumulator charge control. Further, we have composed a microcontroller electronic circuit for the monitoring radiation system. Additionally, an operation algorithm, as well as software for the ATmega328P microcontroller of the Arduino Uno board, have been developed.