Vol 24 No 4 (2025): Informatics and Automation

This paper addresses the problem of coordinated motion planning for multi-link robotic manipulator systems. One of the promising modern approaches to solving this problem is conflict-based planning, which avoids constructing a high-dimensional joint search space by sequentially solving a series of lower-dimensional problems. This is achieved by introducing spatio-temporal constraints whenever conflicts arise in individual manipulator plans, followed by replanning with these constraints in place. Unfortunately, existing methods that use constraints operate with individual time points, which reduces their practical efficiency. In this work, we present a novel conflict-based planning algorithm that utilizes interval-based temporal constraints rather than point-based ones – GECBS-T. Theoretically, the proposed algorithm guarantees bounded sub-optimality of the generated solutions; that is, for any user-defined bound w > 1, the cost of the GECBS-T solution will not exceed w times the cost of the optimal solution. In practice, the proposed algorithm significantly outperforms analogous algorithms in terms of planning speed, as confirmed by experiments conducted in the MuJoCo robotics simulator involving 2–4 KUKA robotic manipulators, each with 7 degrees of freedom.

Controlling a robot swarm with a single remote controller is a challenging task, especially under unstable communication conditions where agents can temporarily lose the control signal, necessitating robust decentralized mechanisms for formation maintenance. This paper presents and tests a semi-centralized control system that enables an operator to coordinate the entire swarm as a unified entity. The system integrates centralized commands from a base station with decentralized position correction via the ESP-NOW protocol. To compare performance in maintaining a rigid formation, the Local Voting Protocol (LVP) and its Accelerated version (ALVP) were applied. Their effectiveness was evaluated in a simulation environment with a group of four drones through experiments involving sharp maneuvers (50° and 75° turns) and significant data packet loss simulations (50% and 80%). The results demonstrate that the Accelerated Local Voting Protocol (ALVP) offers significant advantages over the standard LVP, including faster formation recovery, lower mean positioning error, and greater stability. Specifically, in a series of 20 flight tests with a 50° turn, ALVP successfully maintained the formation in 17 cases, compared to only 3 for LVP, and also showed superior robustness under packet loss conditions. Therefore, the proposed semi-centralized approach using the ALVP protocol is an effective and robust solution for swarm formation control. Future work will focus on conducting physical experiments and integrating obstacle avoidance mechanisms.

The article discusses the development of a mathematical model of human gait for the synthesis of a control system for a mechatronic rehabilitation complex. The relevance of the research is determined by the necessity to create effective rehabilitation technologies for patients with motor function impairments. Existing rehabilitation complexes can be divided into exoskeletons and devices with mechanical linkage (end effectors), with exoskeletons demonstrating higher rehabilitation effectiveness by mimicking natural gait. The scientific novelty of this study lies in the development of a model that takes into account the individual anthropometric parameters of the patient, including body mass and the lengths of limb segments, as well as the ability to simulate foot rotation. Within the framework of the study, a method for dividing the gait cycle into four phases is proposed, each described by a separate system of mathematical equations, which ensures high accuracy in reproducing various stages of movement. To validate the model, a marker-based motion capture system was used, which provided data on movement trajectories. The results showed that the model effectively generates trajectories of sagittal angles of hip, shank, and foot elevation, contributing to improved control of the rehabilitation device. In conclusion, the work emphasizes the importance of mathematical modeling for the development of adaptive control systems that can significantly enhance the rehabilitation process. Further research will focus on refining the model and integrating it with machine learning methods to improve the accuracy and reliability of rehabilitation programs.

This paper proposes an algorithm for short-term rain rate forecasting, RainCast ( Rain Rate Now Cast ), for up to two hours. This area of meteorology, known as 'nowcasting' , is one of the most important tools in many areas of human activity. However, its availability may be severely limited by existing ground infrastructure. In this paper, the authors aim to create a precipitation forecasting algorithm for one such territory using the Asia-Pacific region as an example, based on satellite measurements from the Himawari-8/9 spacecraft. The algorithm combines the advantages of deterministic and statistical approaches to solve the forecasting problem and is based on two neural networks. The first model, a modified version of the physically constrained neural network NowcastNet, generates a preliminary forecast of the general direction of precipitation movement at the mesoscale level. The second model, based on the CasFormer architecture, employs diffusion methods to post-process the initial forecast, refining fine-scale details. The resulting hybrid algorithm, named RainCast, enables short-term precipitation forecasting (up to 2 hours) with high spatiotemporal resolution (2 km, updated every 10 minutes), utilizing solely infrared satellite measurements. Satellite data are converted into precipitation intensity using the algorithm previously developed by the authors. Based on precipitation maps, training, validation, and test datasets were compiled for the algorithm development and forecast quality assessment. The proposed RainCast algorithm was trained on these datasets and compared with other state-of-the-art solutions such as NowcastNet, Casformer, and Earthformer. Analysis of performance metrics demonstrated that the hybrid RainCast algorithm achieves comparable accuracy. For a 2-hour forecast, the Root Mean Square Error (RMSE) was 0.88, the Probability of Detection (POD) was 0.78, the Pearson Correlation Coefficient (PCC) was 0.75, the Structural Similarity Index Measure (SSIM) was 0.91, and the Peak Signal-to-Noise Ratio (PSNR) was 36.63. Visual analysis of the forecasts confirmed that RainCast produces results closest to actual observations, primarily due to the diffusion model's ability to refine fine-scale spatial and temporal precipitation patterns.

The study examines the specifics of training machine learning algorithms on small datasets and addresses the task of forming a training set with high representativeness. It is known that class imbalance in objects, typical for small datasets, negatively affects the performance of algorithms. To mitigate this issue, various data synthesis methods have been developed in machine learning to supplement existing datasets and equalize the number of objects per class. However, these methods do not solve the problem of insufficient representativeness. This article proposes a method for constructing a representative training dataset by specifying the distribution that best corresponds to reality. The distribution is formed for each feature within the informative areas. Informative areas contain characteristic values of features that are most significant for distinguishing classes of objects. The proposed method of constructing areas is based on the idea of gradual expansion, accompanied by an increase in the informativeness of the areas. At the same time, informativeness is understood as a measure reflecting how well objects of different classes can be separated using the considered area. To form a complementary dataset, a generation method has been developed. As a result of its application, the complementary dataset is combined with the original one and forms the specified distribution in the informative area. This distribution can be determined either based on expert knowledge about the subject area, if the true distribution is known, or obtained as a result of computational experiments aimed at finding the most effective option. The applicability of the method is demonstrated by solving the problem of determining the level of temperature anomalies of the mammary glands. It is shown that the considered temperature features are characterized by a normal distribution. Increasing the representativeness of the training set allowed training a classic classification algorithm – logistic regression – with an accuracy comparable to a multilayer neural network. This approach to the formation of a training dataset opens up the possibility of creating more transparent and interpretable artificial intelligence systems.

The article presents a new data augmentation method for measurement systems, designed for industrial equipment condition monitoring tasks. The relevance of the study stems from the significant limitations of traditional synthetic data generation methods, which fail to adequately reproduce complex non-stationary signals with characteristic transient processes, trends, and seasonal variations observed in real industrial environments. The proposed method integrates two advanced techniques: empirical wavelet transform (EWT) and conditional generative adversarial networks (Conditional GAN). The method is implemented in three stages: (1) adaptive decomposition of raw signals into modes using EWT, (2) mode categorization with label assignment, and (3) synthetic data generation using Conditional GAN. A set of statistical metrics was used to comprehensively assess the quality of synthesized signals, including Wasserstein distance (WS), Pearson correlation coefficient (PCC), and root mean square error (RMSE). Experimental studies were conducted on real-world temperature sensor data obtained under non-stationary industrial equipment conditions. The results demonstrate a significant advantage of the proposed method over the traditional TimeGAN approach: a 17% reduction in Wasserstein distance, a 57% increase in Pearson correlation coefficient, and a 21% decrease in RMSE. These findings confirm the method’s effectiveness in reproducing key characteristics of the original signals. The developed method enables the creation of synthetic datasets required for training modern neural network models in industrial equipment diagnostics. Its practical application significantly reduces the costs associated with experimental data collection while ensuring high-quality synthesized signals, as validated by statistical metrics.

The transition to automated regulation systems at unguarded railroad crossings has necessitated the solution of their safety issues. The most rational solution to this problem is the use of video surveillance systems that provide the transmission of images both to the railroad stations, in the area of responsibility of which the railroad crossings are located, and to the locomotives of rolling stock. For this purpose, information transmission systems organized on the basis of broadband and LTE networks are actively used. But since the operation of such networks is organized along the railway tracks, including in conditions of active application of various radiating devices, they are characterized by violation of electromagnetic compatibility conditions, leading to failure of operation as a result of inadvertent blocking of separate channels. Therefore, the analysis of conditions under which the failure of the video transmission network occurs, as well as the predictive calculation of the parameters of radio lines, which provides a given level of stability of the network, is relevant. Technologies and peculiarities of LTE standard networks operation are considered. Indicators and criteria for evaluating the functioning of video transmission lines within the technical capabilities of the standard are substantiated. The mathematical formulation of the research problem is carried out. The initial data for the development of an analytical model of probabilistic assessment of video transmission network functioning are determined. The analytical apparatus for calculating the probability of channel blocking, taking into account the mutual intensity of frequency traffic usage by conflicting devices, is developed. The requirements of GOST R 53111-2008, defining the conditions under which the stability of public communication network operation is ensured, are analyzed. The expression of probabilistic estimation characterizing the probability of disruption (blocking) of network operation determined by both channel noise and fading, and unintentional interference from third-party sources of radio emissions due to violations of electromagnetic compatibility is obtained. The results of analytical modeling are presented, revealing the conditions under which the successful functioning of the video surveillance transmission network is ensured. It is substantiated that the operation of video transmission networks under conditions of mutual interference in violation of electromagnetic compatibility requirements is more sensitive to changes in the ratio of intensity values of network streams and sources of third-party radiations operating in the mode with programmed tuning of the operating frequency than to energy ratios of useful and interfering signals.

This paper presents an innovative approach to clustering and routing in Underwater Wireless Sensor Networks (UWSNs), based on a modified Louvain algorithm that considers sensor distances, the probability of successful message delivery, and the current energy levels of the nodes. The proposed method incorporates a dynamic reclustering mechanism driven by real-time monitoring of energy resources, allowing the network to adapt to environmental changes and sensor status by redistributing roles and restructuring clusters accordingly. The developed algorithm is designed to enhance energy efficiency, minimize data loss, and reduce the number of retransmissions in the context of limited bandwidth in underwater acoustic communication channels. A TDMA-based MAC protocol is also implemented to prevent collisions by assigning independent time slots to sensors, thereby eliminating interference. The approach addresses key resource management challenges in UWSNs by reducing energy consumption, improving data delivery reliability, shortening overall message transmission time, and extending the network’s autonomous operation. The model takes into account the three-dimensional spatial deployment of sensors and optimizes the placement of reference nodes to avoid bottlenecks and excessive energy drain. The primary goal of the study is to construct a network topology that minimizes energy costs and message loss while ensuring efficient routing of data to reference nodes and onward to a mobile sink. The flexibility and adaptability of the proposed solution make it well-suited for real-world underwater applications such as environmental monitoring and ocean exploration.