Abstract. The paper examines approaches to determining the angular position of objects by establishing the direction of propagation of radio waves coming from radio emission sources or formed as a result of reflection from objects. The study is carried out within the framework of grant funding for the scientific and technical project “Development of a hardware and software complex for passive direction finding of radio emission sources of aircraft based on artificial intelligence technology” (contract No. 332/23-25 dated 03.10.2025, project No. AR 327016/0225). Research is also carried out in accordance with the decision of the First Deputy Minister of Defense of the Armed Forces of the Republic of Kazakhstan – Chief of the General Staff of the Armed Forces of the Republic of Kazakhstan dated November 18, 2023, No. 62-3-4068. Within the framework of this area, the Department of Fundamentals of Military Radio Engineering and Electronics conducts a comprehensive study of modern methods for passively determining the direction to radio emission sources. In modern radio direction finding, the accurate and rapid determination of the angular coordinates of radio emission sources, including active jammers, remains a critical task. This requires finding a compromise between measurement accuracy, unambiguous positioning, and equipment noise immunity under complex multipath propagation conditions. This paper conducts a comparative theoretical analysis of amplitude and phase direction-finding methods to systematize their technical limitations and identify ways to overcome them. Using methods of mathematical analysis of radio engineering systems and analytical synthesis of specialized literature, the dependence of equipment sensitivity on the geometry of antenna systems is evaluated. Additionally, numerical modeling of the phase direction-finding method was conducted using the Monte Carlo method by varying the d/λ ratio, signal-to-noise ratio, and signal arrival angle. The analysis shows that the classical amplitude maximum method has low direction-finding sensitivity. Phase methods, conversely, provide higher accuracy; however, as the distance between antennas increases, they encounter the problem of bearing ambiguity. As a result of the study, it is concluded that to achieve the required accuracy and unambiguous measurements, it is most expedient to use combined multichannel approaches. Furthermore, the implementation of modern digital signal processing algorithms (e.g., the MUSIC algorithm) and the use of adaptive antenna arrays can compensate for the impact of destabilizing multipath factors.
Keywords: amplitude direction finding, phase direction finding, detection zone, active jammer, signal, bearing, noise immunity.
Abstract. Not all authentication log events require the same analysis priority; therefore, ranking them using temporal and behavioral context is relevant to information security. This study examines temporal and behavioral features for ranking authentication events by relative risk. The objective is to assess the contribution of behavioral representation to ranking quality and robustness, statistical validity, explainability, and calibration, and to test the technical feasibility of behavioral prioritization on real-world logs. Methods included baseline model comparison, representation ablation, chronological train/validation/test splitting, empirical random-ranking comparison, robustness and statistical analyses, explainability, calibration, low-and-slow temporal-horizon sensitivity, and behavioral feature-group ablation; NDCG@100 was the primary ranking metric.In the synthetic experiment, R2_behavioral_core achieved a higher mean NDCG@100 than R0_minimal (0.126952 vs. 0.071314) and outperformed it in 10 of 12 scenario–model combinations. Across 60 paired observations, mean ΔNDCG@100 was +0.046803 with a 95% confidence interval of [0.035444, 0.057751]. Empirical random-ranking analysis provided an additional reference for interpreting the absolute ranking scores: for Logistic Regression with R2_behavioral_core, NDCG@100 was 0.189052 versus a random mean of 0.010045 in the baseline scenario and 0.145514 versus 0.004943 in the rare-attacks scenario. The behavioral advantage remained condition-dependent and weakened under low-and-slow activity. Explainability and calibration were used to characterize score contributions, explanation stability, calibration quality, and ranking preservation.
For 1,760,511 real-world log events, prioritization scores and unique ranks were calculated from five past-only behavioral features, satisfying all 9/9 temporal/representation integrity requirements. Predictive effectiveness was not evaluated because verified security labels were unavailable. The synthetic results support the incremental ranking value of temporal and behavioral context, while the real-world analysis demonstrates the technical feasibility of implementing the corresponding past-only behavioral prioritization procedure. Confirming predictive effectiveness in operational settings requires temporal external validation using logs with verified security labels and semantic field documentation.
Keywords: authentication events, behavioral analysis, risk ranking, temporal validation, security event prioritization, machine learning, explainability.
Abstract: The work is devoted to the development and software implementation of a specialized web interface for data processing, visualization and predictive analysis of corrosion resistance of protective coatings. The article presents a mathematical apparatus combining the equations of diffusion of aggressive media through polymer matrices (Fick’s second law), the kinetics of electrochemical dissolution of a metal substrate (Tafel equation) and empirical models of degradation of protective layers. Based on the proposed formalisms using Python, the Streamlit framework, and the Pandas, NumPy, and Plotly libraries, an interactive software package was designed. The developed tool makes it possible to automate the input and filtering of experimental data, perform multifactorial regression analysis, build 2D and 3D graphical dependencies, and predict the estimated uptime of coatings. The applied benefit of this work is the formation of a convenient digital twin, which helps to significantly reduce the number of experiments in the process of searching for optimal characteristics of insulating compounds.
Keywords: corrosion resistance, software solution, computer simulation, Python, Streamlit, polymer films, regression methods, graphical data visualization.
Abstract. Pressure-pulsation measurements in aircraft testing depend on the sensing element, connecting passages, housing cavities and supply circuit. This study separates their effects on frequency response, local temperature compensation and uncertainty under explicit assumptions relevant to semiconductor sensors intended for extreme conditions. Thin-plate and second-order models, viscothermal tube transfer matrices and an additive decibel uncertainty model are checked through limiting cases, independent calculations and Monte Carlo propagation. A single bandwidth boundary cannot jointly identify natural frequency and damping. A lower frequency estimate near 195 kHz requires an assumed 65 kHz bandwidth and damping of 0.01–0.10. For a neck 0.8 mm in diameter and 1 mm long, with a 1 mm³ cavity, the complete cascade has a peak at 25.728 kHz and a bandwidth of 7.981 kHz. A 20 mm probe of radius 0.40 mm reduces the bandwidth to 0.981 kHz. Cavity thermal losses and diaphragm acoustic back-loading are neglected. With assumed temperature coefficients and a 6 V supply, the restricted topology provides 13.846 percent of the required local compensation; zeroing the first derivative does not establish stability across the temperature range. Two illustrative budgets give expanded uncertainties of 2.491 and 0.980 dB for a coverage factor of two, and 95 percent interval half-widths of 2.214 and 0.950 dB. The results are conditional computational examples, without new experimental validation or verified accuracy claims for a particular sensor. The work provides a basis for testing the model assumptions and for designing a subsequent measurement protocol.
Keywords: piezoresistive sensor, frequency response, cavity resonance, acoustic probe, parameter identification, temperature compensation, measurement uncertainty, numerical verification.
Abstract. Political communication has migrated to social platforms, where opinions are expressed through short, code-switched, and rhetorically dense messages. Conventional sentiment analysis systems, tuned for product reviews, transfer poorly to this domain because political polarity is entangled with stance, irony, and discourse structure. This article surveys the principal families of sentiment analysis methods—lexicon-based, classical machine learning, deep neural, and transformer-based approaches—and introduces PoliSent, a hybrid discourse-aware architecture for political sentiment analysis on the Internet. PoliSent unifies three complementary signals: multilingual contextual embeddings produced by an XLM-RoBERTa encoder, a domain-specific political lexicon (PoliLex) with valence shifters and negation handling, and a graph-attention module operating over Universal Dependencies parses. A learned gating mechanism fuses these signals, and a multi-task objective jointly predicts sentiment, stance, and rhetorical tactic. We evaluate the framework on PoliWeb-CA, a newly annotated corpus of 18,500 Kazakh- and Russian-language political messages collected from microblogs, messaging channels, and news comment threads. PoliSent attains a macro-F1 of 0.904 and a Matthews correlation coefficient of 0.857, surpassing a fine-tuned KazRoBERTa baseline by 4.6 macro-F1 points and outperforming all lexicon, classical, and deep baselines by a wide margin (p < 0.01, paired bootstrap). Ablation studies confirm that each architectural component contributes measurably, with the dependency-graph attention and lexicon augmentation jointly responsible for a 5.8-point improvement over the contextual backbone. The results demonstrate that explicit linguistic structure remains valuable even in the era of large pretrained encoders, particularly for low-resource and morphologically rich political discourse.
Keywords: sentiment analysis, political discourse, opinion mining, natural language processing, transformer models, graph attention networks, stance detection, low-resource languages, Kazakh; multilingual NLP.
Abstract. The objective of this study is to improve the quality of agricultural engineering for cattle feed production in Eastern Kazakhstan. Kazakhstan’s agricultural production is considered a high-risk area, and therefore, producer and consumer risks of agricultural products are used as key criteria for agricultural engineering quality. Implementing these tasks in this scientific and production context is particularly challenging due to the insufficient scientific study of the subject area, particularly in terms of regional natural, climatic, and geographic aspects. In modern digital production management systems, proactive monitoring and control are crucial. This study explores a new approach that transforms management into an agricultural engineering system, in which risks are considered as criteria for the digital maturity of the business environment. In livestock farming, feed quality plays a crucial role, and monitoring it is impossible without the use of modern digital solutions at all stages of feed production. To qualitatively evaluate control processes based on digital maturity criteria in both integral and differential formats, this paper proposes a new model based on the principles of fuzzy control agents, which improves the quality of decision-making.
Keywords: agricultural engineering, feed production, risk, digital maturity, control, model.
Abstract. Automated object recognition systems are widely used in industry, robotics, and environmental monitoring, with intelligent waste sorting being one of the most promising application areas. Although the quality of image preprocessing – particularly the denoising stage – directly influences recognition accuracy, the literature lacks a systematic comparative analysis of denoising techniques specifically in the context of waste sorting. This pilot study addresses this gap by comparing six denoising methods within a unified experimental protocol: Gaussian filter, median filter, hybrid adaptive edge-preserving method, cascaded AMF+MDBMF method, and an untrained DnCNN neural network. The experimental dataset included 15 representative color images of recyclable objects from three classes (plastic, glass, metal) selected from the Kaggle TrashNet dataset, distorted by three types of artificially added noise: Gaussian noise (σ = 25), salt-and-pepper impulse noise (p = 5%), and Poisson noise. Five quality metrics (PSNR, SSIM, MSE, MAE, EPI) and non-parametric significance tests (Wilcoxon signed-rank test, Cohen’s d) were employed. In addition, the computational complexity of each method was measured. Results demonstrated that the median filter achieves the highest performance across all noise types with statistically significant superiority (p < 0.01). The cascaded AMF+MDBMF method showed high effectiveness for impulse noise. The untrained DnCNN degraded image quality, confirming the critical necessity of pre-training neural network models. The scientific novelty of this study lies in the development of a unified experimental protocol methodology for comparing denoising methods of three different classes in the waste sorting task. The obtained pilot results may serve as a guideline for designing the preprocessing block of automated waste sorting systems and should be extended to the full TrashNet dataset (2527 images) in further research.
Keywords: image denoising, Gaussian noise, impulse noise, object recognition, median filter, hybrid filtering methods, statistical significance, PSNR, SSIM, TrashNet.
Abstract. This study presents a method for improving the safety of autonomous navigation of a mobile robotic platform through multi-sensor data fusion. The main objective is to reliably estimate the robot’s position and generate a safe motion trajectory in the presence of obstacles by integrating data obtained from LiDAR, an RGB camera, an inertial measurement unit (IMU), and wheel encoders. The proposed method integrates the dynamic model of the mobile platform, a multi-sensor data fusion algorithm, an environmental assessment module, safe trajectory planning, and real-time motion control within a unified framework. The dynamic model accounts for the robot’s motion parameters, wheel angular velocities, and control inputs. Based on the fused sensor information, the current position and motion state of the robot are estimated, and the resulting data are transferred to the trajectory planning and control system. During safe trajectory generation, the minimum distance between the robotic platform and surrounding obstacles is considered to determine a motion direction with a reduced risk of collision. The effectiveness of the proposed method was evaluated using MATLAB and ROS–Gazebo simulation environments. The experiments investigated the robot’s motion along a predefined route, detection of static obstacles, and their safe avoidance. In addition, external noise was introduced into the sensor measurements to evaluate the robustness of the proposed approach to measurement errors. The results demonstrated that multi-sensor data fusion improves the stability of robot position estimation, enables consistent trajectory tracking, and contributes to safer autonomous navigation. The proposed method can be applied to the development of autonomous navigation systems for mobile robots in industrial automation, warehouse logistics, and service robotics.
Keywords: mobile robotic platform, safe navigation, multi-sensor data fusion, LiDAR, RGB camera, IMU, wheel encoder, dynamic model, trajectory planning, ROS–Gazebo.
Abstract. In the context of rapid changes in the labor market and the increasing availability of educational data, the development of artificial intelligence (AI)-based career recommendation systems has become an important research direction. However, the influence of data types and algorithm selection on the predictive accuracy and computational efficiency of such systems remains insufficiently systematized. The aim of this study is to systematize current research on AI-driven career recommendation systems and to identify relationships among data types, algorithms, and evaluation metrics. The study is based on a systematic literature review conducted in accordance with the PRISMA 2020 guidelines. The analysis was performed within the “data type–algorithm–evaluation metric” framework. The results show that classification tasks dominate the reviewed studies (56%), while structured tabular data represent the most frequently used data type (76%). The highest predictive performance was achieved primarily by ensemble machine learning methods and neural network models. At the same time, a significant gap was identified regarding the assessment of computational efficiency, as 76% of the analyzed studies did not report relevant computational performance measures. Furthermore, the findings indicate that algorithm selection is strongly associated with the characteristics of the underlying data. The study contributes to the systematization of algorithmic approaches used in career recommendation systems and highlights the need for standardized evaluation protocols, systematic reporting of computational efficiency, and broader adoption of multimodal data in future research. In addition, the development of AI-driven recommendation systems tailored to the specific characteristics of national labor markets and educational environments represents an important direction for future studies.
Keywords: artificial intelligence, career recommendations, recommender systems, data types, machine learning algorithms, predictive accuracy, computational efficiency.
Abstract. The research is devoted to solving the urgent problem of modernization of technological cycles at mining and processing plants. The main focus of the article is on automation of the primary stage of separation of fossil raw materials. Today, in many enterprises, quality control of rock mass is based on visual inspection and is carried out by the operator. It is generally believed that with this approach, the human factor introduces subjectivity into the assessment and reduces the accuracy of impurity fixation. That is why the article explores the possibility of implementing computer vision systems for operational sorting.
The research focuses on the development and testing of a binary classification method for digital images, which makes it possible to effectively separate streams into the target product (coal) and waste rock. In the framework of this work, the Random Forest algorithm was chosen as an architectural solution, the hyperparameters of which were optimized by the lattice search method. During the preliminary tests, the algorithm showed stable results in dusty conditions and changing lighting. To train and test the model, a data set of 4027 images of the mountain range was collected. The experiment was based on a comparative analysis of the proposed method with the methods of convolutional neural network (CNN), logistic regression and decision tree. The results confirmed the potential of this method. The model achieved a classification accuracy of 96.5% with an F1-score of 0.896 and a coal detection completeness of 85.7%. It has been found that with accuracy comparable to convolutional networks, the chosen algorithm has an advantage in resource efficiency and the ability to work on Edge devices without a GPU, providing performance of 30-35 FPS. The research results allow us to conclude that the achieved indicators, as well as the stability of the algorithm, make it possible to successfully integrate it into the monitoring system. The proposed solution can become the basis of an autonomous control system at a mining and processing plant without human intervention.
Keywords: computer vision, machine learning, Random Forest, rock classification, conveyor automation, coal industry.
Abstract. This paper examines an approach to building an event-driven serverless architecture for a distributed information system operating under uneven and peak load conditions. Modern digital services are characterized by sharp fluctuations in the intensity of incoming requests, which requires maintaining operational stability, acceptable response times, and the ability to quickly scale computing resources. Traditional monolithic and container-based solutions in such environments often require upfront capacity reservations or respond to load surges with delays. The goal of this study is to develop and experimentally evaluate an architectural solution in which request processing is organized as a stream of independent events using a serverless computing model. To this end, a formalized event model and input load generation scheme are proposed, enabling the reproduction of both normal operating modes and short-term peak impacts. The effectiveness was assessed based on a series of controlled computational experiments under various load scenarios. The key metrics used were average processing latency, the p95 metric, system throughput, and the error rate during periods of increased activity. The results obtained during the study demonstrate that as the load increases, the increase in latency is manageable, and the system maintains operability during short-term overloads. This allows us to consider the proposed approach as a promising solution for scalable distributed services.
Keywords: event-driven architecture, serverless computing, distributed information systems, scalability, peak load, tail latency, performance.
Abstract. This paper presents an experimental evaluation of the efficiency of the proposed digital signature based on a Verkle tree using the Chinese Remainder Theorem. A software implementation of the algorithms for key generation, signature formation, and verification has been developed. In the proposed scheme, the Verkle tree is used for compact representation of commitments, while the Chinese Remainder Theorem is applied to optimize modular computations and improve the computational efficiency of signing and verification operations. An analysis of the time characteristics of the algorithms was carried out, and complexity indicators were obtained. Experimental results were obtained on a fixed computing platform with multiple test runs to ensure statistical reliability. A comparative analysis was performed with a digital signature based on the classical polynomial commitment scheme Kate–Zaverucha–Goldberg (KZG) in terms of the main signature parameters and execution time. The obtained results demonstrate the potential of using the Verkle tree in combination with the Chinese Remainder Theorem for constructing compact and computationally efficient digital signatures.
Keywords: Verkle tree, vector commitment, polynomial commitment, Chinese Remainder Theorem, digital signature, authentication, verification.