Nataliia Stelmakh

The paper considers the problem of automated determination of the deformation field of a lower-limb residual limb based on contact measurements and matrix processing of experimental data. The relevance of the study is determined by the need to improve the accuracy of individual prosthesis design, since traditional methods for determining residual limb parameters mainly describe its geometric shape but do not fully account for the deformation response of tissues under local loading. An approach is proposed in which the residual limb surface is represented as a discrete set of contact points corresponding to the cells of a matrix measurement system. For each local loading case, a displacement vector of all contact points is formed, and the set of such vectors is combined into an experimental data matrix. After normalization by the applied pressure, this matrix is interpreted as a mutual influence matrix describing the spatially distributed deformation response of the surface. Based on the obtained matrix representation, a method for reconstructing the residual limb geometry and estimating local deformation characteristics is proposed. The developed approach can be used as a mathematical basis for creating automated residual limb analysis systems and for further design of individual prostheses.

This paper addresses the urgent scientific and applied problem of automatic tracking of highly maneuverable Unmanned Aerial Vehicles (UAVs) using systems based on platforms with limited computing power (Edge Computing). The paper analyzes the shortcomings of classical correlation trackers and detectors based on deep neural networks when tracking targets with non-linear trajectories. A hybrid tracking method is proposed, combining the speed of a Kernelized Correlation Filter (KCF) and the accuracy of a neural network detector (YOLO11s). A key feature of the method is the developed algorithm for adaptive Kalman Filter correction, which utilizes a dynamic, scale-invariant Prediction Error metric as a trigger for motion anomaly detection. This allows the system to distinguish between measurement noise and sharp target maneuvers, executing an adaptive state reset using finite differences only at critical moments. Experimental validation on edge hardware (Raspberry Pi 5) using highly dynamic video sequences from the UAV123 and VisDrone datasets demonstrated that the proposed approach maintains an average processing speed of 18.89 FPS. By limiting deep neural network invocations to merely 2.71% of total frames, the algorithm successfully curtails thermal throttling while achieving a global Mean Root Square Error (RMSE) of 259.10 pixels across highly erratic trajectories. The method ensures high tracking reliability without a critical increase in computational load, making it highly suitable for use in autonomous embedded systems.

This paper addresses one of the most challenging kinematic problems in modern visual tracking – the timely detection of unmanned aerial vehicles (UAVs) moving strictly frontally toward the optical sensor, known as the looming effect. It is demonstrated that traditional optical flow algorithms, which calculate object velocity based on lateral displacement on the camera sensor plane, are fundamentally ineffective for such trajectories due to the absence of transverse motion. The goal of this research is the development of optimized mathematical and algorithmic software for computer vision systems capable of detecting head-on collision threats without the use of resource-intensive machine learning methods or heavy convolutional neural networks (CNNs). To achieve this objective, the use of bio-inspired mathematical models is proposed, imitating the functional principles of insect visual systems, specifically the Lobula Giant Movement Detector (LGMD) neuron of locusts and Drosophila neurons (LPLC2). The developed method is based on the analysis of optical flow divergence. It is shown that the critical safety parameter, Time-to-Collision (TTC), can be continuously and accurately estimated as the mathematical ratio of the object's current angular size on the sensor to the rate of expansion of this angular size, regardless of the actual physical dimensions of the UAV or the precise distance to it. The scientific novelty of this work lies in the adaptation and optimization of the lateral inhibition algorithm for hardware platforms with critical size, weight, and power (SWaP) constraints. The developed four-stage video stream processing pipeline – comprising temporal differencing, spatial lateral inhibition, non-linear activation with binarization, and TTC estimation – is implemented exclusively through optimized vectorized spatiotemporal matrix operations. The lateral inhibition mechanism, implemented via a 2D convolution with a specially selected weighting kernel, allows for the effective suppression of global background noise caused by vibrations or camera self-motion, while simultaneously sharply enhancing the useful signal from the rapidly expanding target boundaries. The practical value of the proposed solution consists in enabling high-frequency video stream processing in real-time on budget general-purpose microcomputers like the Raspberry Pi (ARM Cortex architecture) without the need for additional tensor processing units (NPUs) or powerful graphics processors. The hardware implementation of the system allows for the calculation of threat kinematic parameters and the instantaneous generation of logic control triggers via hardware interfaces (GPIO) to activate active safety mechanisms or autonomous evasion systems. Mathematical modeling of the angular size expansion dynamics confirms the reliability of the proposed TTC criterion for differentiating a real collision threat from a steady background.

This study investigates the effect of applied mass loading on the response of a piezoelectric sensor during the registration of mechanical vibrations, enabling an assessment of changes in its sensitivity to vibrational disturbances. An electronic circuit was designed based on a piezoelectric sensor, a signal amplifier, and an ESP32 microcontroller. An experimental study was carried out using various mass loads to determine the sensor’s amplitude response and to establish the dependence of the output signal on the applied load mass.

The possibility of reducing the error in measuring the phase shift of radio signals by the orthogonal method using analogue quadrature demodulators in solving scientific and technical problems in energy, non-destructive testing, research on determining the composition of substances, in radar, radio navigation and radio direction finding is considered. The results of a mathematical analysis of the error in determining the phase shift of radio signals are provided, taking into account the main deviations of the technical characteristics of analogue quadrature demodulators from the ideal mathematical model, which made it possible to evaluate the effectiveness of the proposed improvements. To reduce the error in measuring the phase shift of radio signals by the orthogonal method using analogue quadrature demodulators, it is recommended to use the method of determining the phase shift in the control object at different phase shifts of the query signal applied to the control object. To achieve maximum accuracy, a sufficient number and maximum step of phase shifts of the query signal have been determined.

This paper presents the results of research on the effect of the roller burnishing process on the surface topography and tribological properties of 42CrMo4 steel. Surface topography after burnishing was characterized using parameters defined in the ISO 25178 standard. Tribological tests were carried out on a T-11 tester in a ball-on-disc configuration under both dry and lubricated friction conditions, with variable load, sliding speed, and sliding distance. Each experiment was performed three times. The results showed that roller burnishing significantly reduced surface roughness and produced a texture with dominant valleys, high kurtosis, and strong anisotropy, which favor lubricant retention. Tribological tests confirmed a substantial improvement in wear resistance and a reduction in the coefficient of friction, particularly under lubricated conditions, where wear decreased by more than two orders of magnitude (from 40,026 m to 7,263 m). Optimization of tribological performance using the COPRAS method indicated that low loads (5 N) and sliding speeds of 0.45–0.70 m/s were the most favorable. The findings confirm the high effectiveness of roller burnishing in enhancing the operational properties of steel components.

This paper presents approaches to assessing bone tissue changes before and after osseointegrative prosthetics using modern information and measurement technologies. Particular attention is paid to a comprehensive analysis of the state of bone structures at different stages of osseointegration, which allows for a deeper understanding of the processes of tissue remodelling under the influence of implantation. One of the key research tools is the use of high-resolution computed tomography data. Based on the obtained tomographic slices, three-dimensional (3D) models of bone structures are built using specialised software. These models provide a detailed visualisation of the anatomical features, spatial location of the implant and surrounding tissues, and allow for the analysis of morphological changes resulting from the osseointegration process. In addition to the morphological assessment, mechanical testing of the bone tissue is carried out as part of the work. Loading methods are used to determine physical and mechanical characteristics such as strength, stiffness and elastic properties of the bone in the area of contact with the implant. This allows not only to detect structural changes but also to assess their impact on the functional capacity of the musculoskeletal system after prosthetics. The results are combined into a single data set for comprehensive analysis. This approach allows us to objectively assess the effectiveness of the osseointegration process, identify potential risk areas for implant stability, and formulate recommendations for individualising prosthetic planning for each patient. An important feature of the applied methodology is the possibility of early detection of pathological changes or unsatisfactory implant integration by monitoring the dynamics of morphometric and mechanical parameters in the postoperative period. This opens up prospects for improving the overall success of osseointegrative prosthetics and reducing the number of complications associated with implant instability or rejection. Thus, the developed approach demonstrates high informative value and practical significance, providing a multilevel assessment of bone tissue condition, which is extremely important for ensuring long-term stability and functionality of prostheses in clinical practice.

This article focuses on improving object recognition algorithms for rescue drones, in particular, enhancing the methodology for classifying human poses by expanding the set of key body points and using more effective mathematical models. A methodology is proposed that analyzes 11 key body points, enabling the classification of human positions (standing, lying down, sitting, kneeling, bent) with greater accuracy. Additionally, a gesture recognition algorithm is proposed, detecting gestures such as arm-waving as a signal for help, which increases the effectiveness of rescue operations. The paper also considers the possibilities of implementing the system on the limited hardware resources of onboard UAV computers. Using geometric relationships between body points reduces computational costs without sacrificing accuracy, making the proposed model suitable for real-world use. The conducted research confirms that the improved system can automatically assess victims’ conditions, prioritize rescue efforts, and optimize drone navigation. In future work, it is planned to integrate the algorithms with drones’ multisensory systems and test them in real-world conditions.

This study presents the development of a functional block diagram of a temperature control system for an injection molding machine designed for manufacturing balls from a mixture of polyamide 6 with 30% glass fiber (PA6 GF30). The objective of the research was to create a simulation model that ensures precise temperature regulation during the molding process, taking into account the specific properties of PA6 GF30 materials, such as hygroscopicity and thermal characteristics. The model, developed in the MATLAB software package for numerical analysis, employs a PI controller to achieve stable temperature regulation. Parameter optimization was performed to ensure high-quality ball molding, resulting in temperature stabilization at 180–182°C with an initial drop to 100°C, which corresponds to real-world process conditions. The obtained results confirm the system's suitability for practical use and its potential for further improvement.