The paper proposes an algorithm for decomposing regions of a finite element model (FEM) for the integration of unified cellular structures in the design of lightweight parts. In this algorithm, the considered FEM region is represented as a graph for the purpose of decomposition. Based on this graph, the procedure for identifying break points and the criteria for splitting FEM regions at narrowed areas are described. The use of Dijkstra’s algorithm is justified for determining the weights of vertices located in constricted regions of the FEM. An example of the algorithm’s operation is provided, demonstrating its applicability in the design of lightweight parts with cellular structures.

The purpose of this study is to reveal the scope of research conducted in the fields of artificial intelligence and cybersecurity within the healthcare sector and to analyze the prominent thematic trends in these studies. In the existing literature, no study has been found that comprehensively analyzes research on artificial intelligence and cybersecurity in healthcare using a holistic method. Accordingly, in order to address the research question, “How are the scope and thematic trends of studies focusing on artificial intelligence and cybersecurity in healthcare shaped in the literature?”, a total of 565 articles published in the Scopus database between 2020 and 2025 were examined using the descriptive content analysis method, one of the qualitative analysis techniques. The study reveals that the highest number of publications worldwide originates from Asia, while the most citations are attributed to publications by researchers in the United States and Australia. Furthermore, the research shows that the publications on artificial intelligence and cybersecurity fall into three distinct thematic clusters.

The problem of identifying parameters and simultaneously estimating phase coordinates of a simplified mathematical model of the planar motion of an unmanned aerial vehicle (UAV), defined by stochastic equations including random functions, is solved. The input signal is the elevator deflection angle, and the output signal is the UAV pitch angle, obtained as a result of experimental studies of the UAV flight, and as a result of the structural and parametric identification of the deterministic mathematical model of the UAV. The equations of shaping filters that model the influence of atmospheric turbulence on the angular orientation of the UAV are considered as random functions. The corresponding transformations of the mathematical model are performed, allowing the vector-matrix equation of the model to be represented in linear form. For this mathematical formulation of the problem, the choice of a method of simultaneous estimation and parametric identification based on state vector expansion and the application of the Kalman filtering algorithm is justified. The results of computer modeling confirmed the validity of the application of the considered stochastic model and identification method. This approach can be used to estimate and identify more complex stochastic mathematical models of control systems for technical objects.

Addressing the issues of sudden range reduction and lifespan degradation in agricultural drone lithiumion batteries due to temperature runaway during high-temperature operations, this paper proposes an optimization method for an electric drive battery power supply system. The study initially constructs a battery thermoelectric coupling model to analyze the dynamic relationship between power demand and temperature rise during typical spraying operations. Subsequently, an intelligent thermal management system is designed, which combines active air cooling and phase change material (PCM) cooling strategies and is controlled by an innovative dynamic threshold algorithm. To further enhance system performance, the discharge current curve is optimized using the Particle Swarm Optimization (PSO) algorithm, achieving a better balance between power output and thermal generation. Experimental verification under 35°C environmental conditions shows that the battery peak temperature decreases from 58°C to 49°C, representing a % reduction in the maximum operating temperature. Field tests indicate that the system extends the durability of a single operation by 14 %, while significantly improving the service life, with a capacity retention rate of 92.3 % after 100 charge-discharge cycles, compared to 85.1 % for traditional systems. These results demonstrate that the proposed method effectively addresses the thermal challenges of drone batteries by coordinating and optimizing cooling strategies and discharge management, providing a practical, cost-effective solution for maintaining battery performance in hightemperature agricultural applications, while extending service life and operational reliability.

The aim of the work is to improve the proportional integral differential (PID) controlled system robustness using the discrete adaptive control rule. In the recent decades, the tracking control of the industrial robots’actuators remains an actual problem. In particular, the robust output PID-control stands of significant importance. It is well known, that PID output control provide robustness for the low order plant model. Otherwise, adaptation-based control can be helpful. The criteria of the adaptation are proposed and the gradient method based adaptation rule is evaluated. For the fast adaptation, the control low offered with the second derivative, which assures the dynamics robust quality. On other hand, with the second derivative in control law, it is possible to reach high robustness without adaptation. Hence, in the system with PID control, two alternatives must be compared. Firstly, the not adaptive PID2 controller with second derivative in control low. Secondly, the adaptive PID controller, which fastness achieved with the second derivative in the control law. The discrete adaptation and its comparison with not adaptive discrete PID2 control are analyzed by means of simulation for nominal and perturbed plant. The simulation results demonstrate the high efficiency of the proposed method of adaptation. The proposed criteria of adaptation and structure of the fast adaptive control rule provide stability and the robustness of dynamics, suggested by simulation. The comparison of the systems with PID2 controller and the adaptive one demonstrate advantage of the adaptive controller because of the width stability range. The method is useful only for output control when the output value is measured by sensor. That can be tracking control systems, for instance, robots’ end effectors.

The work suggests complexifying actors within the framework of TD3 which involves the usage of different state vectors for actors and critics in order to assure convergence of the algorithm. It also describes a process of aggregating models, separately trained on datasets or in simulation on tasks with increasing difficulty, stitching everything together step by step into a single end-to-end system. It allows utilizing existing algorithms, such as YOLO, in reinforcement learning systems, performing sensor fusion and gradually adding functionality without losing convergence. Assistance providing allows training systems in simulation from hardcoded algorithms that use simplified states. These techniques are demonstrated on a particular task of building an anti-drone system for armored vehicles.

To date, the mechanical strength testing of electrical apparatus in electrical installations has only been carried out taking into account the static action of the conductors in the event of a short circuit (short-circuit). In reality, short-term dynamic forces act on the conductors during short-circuit faults, which are further transmitted to the apparatuses, gantries and supporting insulators, in which case dangerous mechanical stresses arise, which can lead to their deformation or complete destruction. Based on the above, we can conclude that for a reliable determination of dynamic loads on the main structural elements of open switchgear (OSD) requires the formulation of a dynamic problem, which would involve the consideration of wires and apparatus as a single oscillating system at short-circuit. There are known cases when structural elements of electrical installations, having a large safety margin for static load, were destroyed under the action of relatively small periodically acting forces.

The subject of research is to evaluate the use of neural networks in the university’s educational process. The purpose of the article is to evaluate the use of ready‒made neural networks in the organization of the educational process using factorial, regression and correlation analyses. The main aspects of the use of neural networks, their impact on student academic performance and the effectiveness of educational programs are considered. The use of ready-made neural networks in the university’s educational process has significant potential to improve learning efficiency. The key factors for successful implementation are the technical equipment of the university, the qualifications of teachers and the availability of ready-made solutions. Regression models have confirmed the positive impact of neural networks on student academic performance, and correlation analysis has revealed a strong link between their use and student motivation. It is recommended to: increase the number of hours allocated to the study of neural networks; conduct regular trainings for teachers.

The paper proposes a comprehensive approach to solving the problem of the quality of higher education at the level of the work of the admissions committee of the university: support for the decision-making of the applicant in the process of choosing specialties at the stage of forming the application and the distribution of applicants who have passed the competitive selection and are admitted, according to student study groups, taking into account the level of motivation of those distributed.