Problems of Informatization and Management

Exploratory data analysis and visualization on the example of an e-commerce enterprise

S.R. Suleimanova — 2024-11-04

This article presents an approach to exploratory data analysis and visualization using the example of an e-commerce company. The study examines key stages of exploratory data analysis, including data preprocessing, visualization, anomaly detection, correlation analysis, and cluster analysis, aimed at preparing data for solving machine learning tasks in future research. These tasks include estimating a time series model, identifying trends, seasonal and cyclical components of time series, customer clustering, new customer classification, and predicting the quantity of items sold within customer clusters. The proposed approach can be applied to the analysis of other e-commerce datasets.

Increasing the sensitivity of the sensor of the coordinate machine

O.V. Chalyi — 2024-11-04

The article investigates the sensitivity of the coordinate measuring machine sensor to external factors. Mathematical models have been developed for the evaluation and compensation of systematic and random errors arising from temperature changes and vibrations. A comprehensive model describing the sensor behavior in three-dimensional space, taking into account rotational errors, is presented. Methods for optimizing error compensation based on minimizing the root mean square error are proposed. The obtained results have practical significance for improving the accuracy of coordinate measurements in various industries.

Improving measurement accuracy of aviation parts on a coordinate measuring machine

S.F. Filonenko — 2024-11-04

This article is devoted to the development of methods for enhancing the accuracy of aviation parts measurements on coordinate measuring machines, which is critically important for ensuring quality and safety in aviation manufacturing. A mathematical model is proposed that accounts for the main types of errors: geometric, kinematic, and thermal. An automatic correction algorithm is also developed, allowing for the minimization of these errors' impact on measurement results. The application of this approach contributes to the improvement of accuracy and reliability in controlling geometric parameters, which is essential for the production of structurally complex and critical aviation components.

Method of synthesis of digital controller for UAV automatic control system

O.M. Tachinina — 2024-11-04

In this article the method for synthesizing a digital controller for an unmanned aerial vehicle (UAV) automatic control system is proposed. The suggested method is simple to use and enables rapid parametric tuning of the digital controller for a specific loop of the UAV automatic control system. The synthesized digital controller can be integrated into the software or hardware of the UAV automatic control system. The proposed method can be used to design digital controllers for automatic control systems of any type of UAV.

The method of adaptive K-path routing in special purpose wireless networks

A.A. Skrypnychenko — 2024-11-04

The task of ensuring the end-to-end quality of service of heterogeneous wireless networks (HWNs) with random multiple access is considered. The processes of the emergence of conflicts between the universality and complexity of route selection and resource allocation in BPM are analyzed. A technique for eliminating conflicts between fairness and optimality of resource allocation with different numbers K of routes has been developed. It is shown that with a rational choice of the number of routes, depending on the state and parameters of the autonomous segments and the network as a whole, the functions of each level and, accordingly, the information parameters of the traffic at this level can be attributed to one of two alternative groups: functions that depend on a specific technical implementation networks or application-oriented functions. Accordingly, a global approach to the optimal selection of a radio access technology (RAT) to each of the routes included in the set K is formulated, and an approach to RAT selection using a distributed adaptive search algorithm based on the peak rate at which users are able to receive and transmit is proposed traffic flows. The advantage of the adaptive K-path routing method compared to the classical method was estimated by means of calculations and computer simulations. A distributed system for choosing the number of routes has been developed based on the optimization of the set of routes with the elimination of redundancy.

Method of modular multiplication by invariable number for open key cryptography quick implementation in IoT

V.L. Selivanov — 2024-11-04

The article proposes a method of accelerating the operation of modular multiplication important for cryptographic applications due to use of precomputations. The technology for forming a table of precomputations, whose depend on a invariable multiplier and modulus, has been developed in detail. A formalized description of the proposed procedure for accelerated modular multiplication using precomputations is given. The statement is illustrated by a numerical example. It has been theoretically proven and experimentally confirmed that by using preccomputations according to the proposed method, the implementation of this operation is accelerated by four times. At the same time, the execution time of the basic operation of public-key cryptography – modular exponentiation – is reduced by 1.6 times.

Modernization of flight information reading and processing means

A.S. Savchenko — 2024-11-04

In modern conditions, the task of quickly transferring primary data directly from the on-board recorder of the aircraft to a solid-state information carrier for further processing arises. The article proposes the modernization of flight information reading and processing complexes using the developed prototype of the hardware and software part of the BIF-4T interface block and its own software. This will contribute to increasing the efficiency and comfort of the work of aircraft maintenance specialists, as well as saving time resources and reducing the human factor.

Technology of analyzing electronic media to identify informational and psychological influence

M.O. Riabyi — 2024-11-04

Our many years of research into media information flows have allowed us to reveal a general scheme for implementing informational and psychological influence. The structures created by the decision-making centers (DMCs – “Kremlin towers”) of the aggressor during the hybrid war are designed to conduct information and special informational-psychological operations (SIPO) that support military actions through acts of informational-psychological influence. These operations are implemented by publishing “toxic” messages in news media controlled by the aggressor. The content of these messages poses an informational threat. These messages are then spread further by the aggressor’s network of “bot farms” on social media, simulating public opinion. During wartime, SIPO aim to exhaust society, undermine the morale of the military, and prompt ill-considered military-political decisions by state authorities. Together, structured messages and topics form an “information matrix” that characterizes the information content of individual IPI in the process of conducting hybrid SIPO. The work is devoted to the further development of models and methods of detection and assessment of PR influence to identify and establish the informational content of informational and psychological influence. The refined technology is formulated in an analytical form. The main elements necessary for its practical implementation are identified. The technological process of analyzing electronic media to identify and establish the information content of IPI is described using a UML usage diagram. An example of fragments of an automatically generated report, built using the analyst's AW software, is provided, which form the basis of the final analytical report, and the analyst's conclusion on narratives and identified IPI for the reporting period in the form of abstracts.

Overview of data collection and analysis methods in intelligent information processing systems

M.O. Ryabyy — 2024-11-04

This article examines modern methods of data collection and analysis in intelligent information processing (IIP) systems. A review of available services, their features, advantages, and disadvantages across various fields such as marketing, jurisprudence, medicine, and military affairs is provided. Special attention is paid to the integration of data from messengers and public channels. The article proposes an approach that combines traditional data collection services with sources such as Telegram, Facebook, and YouTube to build a more comprehensive and representative informational framework.

The results of experimental research demonstrate that even non-standard information sources can be effectively integrated into analytical systems for automated detection of patterns, trends, and anomalies. A conceptual model is proposed, which integrates hyper-automation with IIP to improve the efficiency of data collection, processing, and analysis. Experimental research confirms the feasibility of implementing such solutions, particularly for detecting disinformation, forecasting threats, and automating routine tasks.

The article highlights the prospects for developing hyper-automated systems in the context of a rapidly evolving information environment.

Development of communication strategies for small agricultural enterprises: online platform creation

N.S. Vovk — 2024-11-04

An analysis of current trends shows that the development of information technology plays an important role in agriculture, helping to improve the efficiency and competitiveness of the industry. In this regard, there is a need to improve communication systems with consumers to ensure effective interaction and expand market opportunities for agricultural enterprises.

The article discusses the role of an online platform in the development of small-scale agriculture, in particular through the creation of a website based on a content management system. It emphasizes the potential of this tool to increase consumer confidence, improve communication with them, and attract new customers. The advantages of using a CMS are considered in the context of ease of use, flexibility in customization and the ability to quickly update information about products and services.

Based on the analysis of a family farm, the authors demonstrate how to effectively implement these technologies to ensure successful customer interaction. In addition, the authors highlight the process of building a website, taking into account the needs and expectations of the target audience. This approach allows not only to improve communication with consumers but also to create a solid foundation for the development of agriculture in the online environment.

Using the Levenshtein algorithm for the categorization of information systems users

Y.B. Artamonov — 2024-11-04

The study is devoted to the application of the Levenshtein distance algorithm to categorize users based on behavior in information systems, with an emphasis on dynamic adaptation in e-learning environments. The experiment used data from the MS SQL Server online course and the hardware and software complex for teaching the Braille font, which allows you to switch the content based on scenarios that correspond to the category of users. The results show the effectiveness of using the Levenshtein distance algorithm for user categorization and training has maintained a positive trend of advancement according to predefined training scenarios. The study also highlights the potential of adaptive e-learning systems to improve learning outcomes by adapting content to the individual needs of users.

Internet network model with account of network location

S.S. Chumachenko — 2024-07-01

The purpose of this research is to prove the usefulness of the conclusions of the percolation theory for the assessment of the connectivity of the Internet of Things network. The probability that there is a route between two pairs of nodes determines the connectivity of the network. When this probability approaches unity, the network is considered connected. For example, finding the shortest paths between all pairs of nodes (the vertices of the graph describing the network) can be used to assess connectivity in a particular situation. Two reasons make this problem difficult to solve. First, the computational resource required for this grows proportionally to n3, where n is the number of vertices of the graph. Second, existing analytical approaches are usually focused on specific cases rather than on a large number of nodes.

The article proposes a network model in the form of a lattice, which serves as a logical basis for structuring connections between nodes located in different rooms. The shape of the grid is controlled by the layout of the building. The obtained model makes it possible to evaluate the efficiency of the communication channel using the estimation of the PSP and the general ratio of the SS for the nodes of the network. The use of the model allows us to confidently state that in a modern multi-story structure, if the nodes are strategically placed in each room, it is quite possible to create a wireless communication network. In addition, the quality of the radio channel allows you to use data transfer rates that meet the highest standards of modern technologies.

Priority management in IT service for airline employees

V.F. Suraiev — 2024-07-01

The purpose of the processing is the automatic prioritization of IT service requests for employees of airline`s departments. The object types affecting request priorities identified. The simple function for calculating the priorities of requests based on the analysis of such components as the user's department, position, the location of the problem, the service that must be restored, the work that will solve the problem and the plan for performing this work proposed to use. The result is obtained by successive transformation of relations of sets departments, positions, locations, services, works and their execution plans.

Overfitting in machine learning

M.V. Struk — 2024-07-01

The problem of overfitting in machine learning is relevant and important for achieving high accuracy and reliability of predictions on real data. This article is dedicated to exploring the problem of overfitting from a mathematical perspective. It begins with a general overview of the problem and its importance for scientific and practical tasks such as pattern recognition, forecasting, and diagnostics. Starting with defining key concepts such as model complexity, sample size, bias-variance tradeoff, and dispersion, the text reveals the relationship between them and the influence of sample size on the learning process. To demonstrate these concepts, Python code is developed that uses polynomial regression as a model for analysis. Through the creation of synthetic data and fitting different models to them, the phenomenon of overfitting and its impact on prediction accuracy is illustrated. The concluding remarks emphasize the importance of understanding the mathematical aspects of overfitting for developing reliable and effective models in machine learning. Further analysis of recent research and publications in this field demonstrates various approaches to solving the problem, including regularization methods, the use of ensemble methods, and the development of new neural network architectures. Unresolved aspects, such as finding the optimal balance between model complexity and generality, are highlighted for further investigation. The ultimate goal of the article is to identify the key aspects of the overfitting problem and formulate goals for further research in this area.

A method for prioritizing IT incidents at critical information infrastructure facilities of the state

V.M. Sydorenko — 2024-07-01

The article is devoted to the development of a method for prioritizing IT incidents at critical information infrastructure facilities of the state. The study considers the main approaches to classifying and managing IT incidents, such as ITIL, COBIT, ISO/IEC 20000 and NIST Cybersecurity Framework. The proposed method is based on the use of the pairwise comparison method (AHP) to assess and prioritize threats, taking into account their impact on different levels of critical facilities. The article describes in detail the stages of method development, including threat identification, local and global prioritization, and synthesis of the results obtained for effective IT security management. The proposed approach allows rational allocation of resources, ensuring the reliability and resilience of critical information infrastructure. Experimental studies confirm the practical value of the method, which makes it a useful tool for increasing the level of security and effective response to IT incidents in the face of modern threats to IT-security.

Method of distributed modular exponentiation on IoT terminal microcontrollers with protected involvement of cloud computing

O.V. Rusanova — 2024-07-01

The method of distributed modular exponentiation on IoT terminal microcontrollers with secure involvement of remote computer systems is theoretically justified, developed and researched, the use of which allows to accelerate the implementation of this basic operation of public key cryptography.

Acceleration is achieved by reducing the computing load on the terminal computing platform. The paper theoretically substantiates this possibility by increasing the load on cloud computing, provides a detailed description of the method, the work of which is illustrated by a numerical example, and presents a detailed assessment of efficiency.

Theoretically and experimentally shown. that the proposed method allows to speed up modular exponentiation by 3-4 times compared to known methods of its implementation on IoT terminal devices with the involvement of cloud computing.

A model for identifying the source of pseudorandom number sequences based on a hybrid neural network

D.P. Proskurin — 2024-07-01

This article presents a model for identifying random number sources based on a hybrid neural network. The proposed model combines recurrent (RNN) and convolutional (CNN) neural networks to achieve high classification accuracy. The study discusses the key stages of model development, including data preparation, model construction, training, and performance evaluation. Experimental results confirm that the model can effectively identify random number sources with an accuracy of more than 95% for some generators. The developed approach provides high reliability and can be applied in various fields, including cryptography and modeling.

One-directional mapping for cryptographic security of the quasi-regular language sentences to interact light aviation systems subjects

M.K. Pechurin — 2024-07-01

One from the problems of the interaction processes (communications) protection between subjects of ultralight aviation systems that functioning (perform targeting) in an autonomous mode is considered – the problem of communications cryptographic protection in conditions when the interacting components information and communication capabilities are low An asymmetric cryptographic protection system based on the use of a modified regular language of communications is proposed, relevant to the low availability of information and computing resources.

Algorithm for calculating the power system voltage frequency using quadrature components

R.V. Petrosian — 2024-07-01

The scientific article solves an important scientific and technical problem – minimizing the error in calculating the power system voltage frequency. The calculation of the voltage frequency is important because it helps to ensure the reliability and stability of power facilities. Long-term operation of equipment with a significant deviation of the voltage frequency from the nominal value leads to increased wear and premature failure. In addition, a mismatch between the frequency and the nominal value can lead to additional errors when measuring other parameters of power facilities.

The algorithm is based on calculating the power system voltage frequency using quadrature components. In the course of the work, a structural scheme for calculating the voltage frequency using digital filters was proposed. The synthesis of digital filters was also performed. To verify the results obtained, a computer simulation was performed using the Python programming language. An algorithm for calculating the voltage frequency in the power grid is proposed.

Construction of the software of the information system for the formation of specialized thematic virtual communities

R.S. Pazderska — 2024-07-01

The article is devoted to the study of the software operation processes of specialized thematic virtual communities based on informational messages and user behavior in the social network. Attention is also paid to the research of virtual communities, which led to the emergence of specialized thematic virtual communities.

The results of the information system for the formation of specialized thematic virtual communities are presented, which will improve the process of dissemination of information content by users. Attention is focused on the following processes: creation of a specialized thematic virtual community based on the selection of users, analysis of the activity of an existing specialized virtual community and its search for the purpose of publishing the specified information content.

A formal modeling for the subject area is presented, on the basis of which the methodology formulation of the degree of correspondence of the information content of the specialization and the topic according to the defined keywords and tags, which were used to implement the system's tasks, took place.

The research results can be used to optimize the work of specialized thematic virtual communities.

Soft-adaptation of reference algorithm for discrete control

V.S. Nozdrenkov — 2024-07-01

An applied technique for quickly transitioning from a pre-optimized reference control algorithm to its discretized (software) alternative for the needs of discrete control of objects operating on the “on-off” principle is proposed. This technique can effectively facilitate the design and implementation of low-cost automatic control systems, in particular, based on a unified information and control network organized by the client-server principle. The method application involves some stages, namely: basic formalization of the control circuits and obtaining of optimal transient processes through modeling (implemented using the MatLab Simulink environment in the current article), identifying the optimal reference control algorithm and its parameters, an adaptation characteristic constructing, software implementation of the adaptation characteristic on controller level.

Using a resource classifier in traffic forecasting models for SDN networks

Y.O. Kulakov — 2024-07-01

This publication reviews existing solutions for predicting SDN traffic. The proposed solution uses a resource classifier of client requests for predicting the network response to these requests. This short-term forecasting is performed using a convolutional neural network. The article considers aspects of parallel computing and forecasting in a dynamically reconfigurable system. Forecasting results can be used in QoS/QoE services, SDN controller for traffic routing and detection of anomalous traffic by network protection systems.

Load balancing method in SDN networks using artificial intelligence

Y.O. Kulakov — 2024-07-01

This paper provides an overview of load balancing in corporate networks using SDN technology and artificial intelligence. It discusses the features and methods of using AI in these networks and identifies potential problems.

The paper reviews the method of load balancing in corporate SDN networks using AI and concludes that AI can automate the process of load balancing in the network. The implementation of AI can enhance network efficiency, decrease costs, and bolster security. This paper suggests the use of artificial intelligence in corporate networks through SDN technology to improve network efficiency and security.

This paper suggests the use of artificial intelligence in corporate networks through SDN technology to improve network efficiency and security. The paper reviews the method of applying AI in these networks and outlines promising areas of research.

This paper presents a theoretical overview of load balancing in enterprise SDN networks, considering their features, and establishes the foundation for future research in this area. The integration of artificial intelligence into corporate networks is a pressing task as it enhances network efficiency and security while also creating new opportunities for development.

Control of linear motors with gas lubrication of the coordinate measurement system

V.P. Kvasnikov — 2024-07-01

The problem of the influence of non-Gaussian disturbance on the control system during the measurement of coordinate measuring system of mechanical quantities of objects with a complex spatial surface is described in the article. It is shown that the accuracy of the filtering of the random process of measuring mechanical quantities is significantly affected if to take into account the disturbing factors with non-Gaussian probability distribution density.

Challenges of data collection for monitoring technical characteristics and location of specialized airport vehicles

Y.Y. Karpov — 2024-07-01

The study analyzed the procurement of Boryspil Airport from open sources, as well as regulatory and legislative documents regarding data transmission via radio modems. It was found that over the past few years, the airport has been tendering for systems using GPRS cellular network operator networks (networks of third-party companies not controlled by airport services). The aim of this study is to calculate the parameters of the radio network, determine the data to be transmitted from sensors to the database, select radio modules to provide coverage of the airport's radio network area, create an architecture of a computer-integrated monitoring system for technical characteristics and location of specialized airport vehicles with the goal of transitioning from using networks of third-party companies to an independent internal network. The study calculated the parameters of the airport's radio network, analyzed the prospects for the development of the data transmission subsystem, additional possibilities for determining the location of special vehicles using radio modules and GPS. It considered reducing the impact of other unwanted sources of interference on the radio module's operating frequency, reducing the risk of data interception and decryption transmitted by the radio network, and choosing a radio module based on SX1280 for the development of a prototype for further research.

Using an object tracking algorithm by video image for the implementation of an autonomous target tracking function for UAV

I.A. Zhukov — 2024-07-01

The article examines the spheres of UAV commissioning. The problem of controlling the UAV by the operator through the control panel is described. The paper proposes an approach for implementing the autonomous function of following a given object. A software part consisting of a video stream processing module, object tracking, and control command generation was developed. The program additionally implements mechanisms to increase system reliability, such as data processing to prevent sudden movements and an algorithm to restore tracking when an object is lost. Further steps for the integration of the MAVLink library for the formation of control commands and the assembly of the hardware part of the system based on a Raspberry Pi single-board computer and a Pixhawk flight controller are described.

Method for cryptographically strict identification with password reuse blocking

O.A. Verba — 2024-07-01

A method for cryptographically strict identification of remote users with the combined use of irreversible transformations of modular arithmetic and functionally interconnected session passwords is theoretically justified and developed, which eliminates the possibility of their reuse by an attacker.

In addition, to speed up the identification process on the system side, it is proposed to use irreversible transformations of modular algebra that are asymmetric in the volume of calculations.

A mathematical justification for the proposed method and numerical examples that illustrate its operation are given. It has been theoretically and experimentally proven that the proposed method makes it possible to block the reuse of user passwords, as well as speed up the process of identification by the system by 2-3 orders of magnitude.

Multi-model data representation for ontology-driven decision-making

Y.P. Chaplinskyy — 2024-04-01

Nowadays ontology-driven decision-making is the result of the combination and integration of knowledge and data of different nature and purpose, solving multiple interconnected decision-making problems, synthesizing different points of view on the problem and the decision-making process. For many decision-making problems, the use of only one data model, usually a relational one, is not enough. Today, to solve the problem of "diversity" of data, the concept of multi-model data management is used. Modern aspects of decision-making based on the use of ontologies as a tool of understanding and presentation of decision-making areas and processes, which integrates the methods of system, process and situational analysis are discusses. The possible types of data representations (models) that are necessary for the implementation of ontology-guided decision-making, and also defines the corresponding types of information are defined. In the work, polyglot persistence storage, multi-model storage and cloud-based storage were considered and analyzed for the implementation of multi-model data management. The conclusion is made about the choice of multi-variant storage as a means of multi-model presentation of data within the framework of projects carried out by the authors of the article. The selected software tools for implementing the defined multi-model data representation are described.

Synergy of cosine similarity and generative artificial intelligence

T.V. Kholyavkina — 2024-04-01

First and foremost, the synergy between cosine similarity and generative artificial intelligence in the context of developing a contextual search application is a promising direction, especially given that most search systems rely on keyword-based search. The new approach proposed by artificial intelligence is based on neural and semantic search, focusing not on individual keywords but on their context and relationships, which opens up broad possibilities for improving the accuracy and relevance of search results.

To meet users' needs, chatbots are being developed. Chatbots are utilized in various fields, including customer service, consultation, user support, and other tasks that require interaction with people through text or voice interfaces. Their popularity is constantly increasing in response to demand from users, the majority of whom regularly use chatbots in their everyday lives. The article covers concepts of artificial intelligence, AI-based chatbots, contextual search, and cosine similarity. The focus is on the application of contextual search based on artificial intelligence. An explanation of the approach and specific details of the implementation of its development is provided.

A model of an opportunistic computer network control system

O.V. Tolstikova — 2024-04-01

The mobile multi-service network is considered – a segment of the information and communication system of control, management, and maintenance of a large air hub. It is shown that under conditions of sudden entry of aircraft into and out of the network coverage area, the latter acquires the properties of an opportunistic network, and sometimes – a network with tolerance to delays and disconnections. It has been established that as a result of a change in the number of active network nodes, the network topology may change and there is a need to change data delivery routes. A model of the processes observed in the opportunistic network is built. As a mathematical basis of the model, a non-homogeneous alternating restoration process is used, which takes into account changes in the transition intervals from one phase state to another. It is proved that with different variants of dominance, the appearance of new mobile nodes or the disappearance of nodes that were part of the network earlier, the asymptotic convergence of the network states changes accordingly.

Method of project management based on Story Points estimations

O.V. Rusanova — 2024-04-01

The paper is devoted to the development of building project management systems methods that allows to perform automatic project scheduling, where tasks are estimated in story points. The paper reviewed the existing systems and proposed an approach to the automation of project scheduling in minimum time, and using the minimum employees number who will complete the project in the given time.

The proposed approach allows taking into account the velocity, specialization and salaries of employees. Taking into account these features, and using of the tasks complexity in story points, brings scheduling closer to reality, which allows to increase the accuracy of scheduling, reduce the time of project execution and/or costs for employees and the project in general.

ІЕЕЕ 802.15.1 МАС-to-PHYSICAL level transition protocol and one-directional parsing function

M.K. Pechurin — 2024-04-01

The procedure of inter-level transformations of PDUs is considered on the example of the MAC sub-level of the channel and physical levels of the Reference model of the interconnection of open systems for interacting UAVs of the first class, taking into account the main recommendations of IEEE 802.15.1 regarding the protection of information from unauthorized access. Distinctive features of first-class UAV computer systems and networks are low transmission speed and limited computing power of computer equipment. The recommendations regarding the relationship between the channel and physical layers, subject to the application of the IEEE 802.15.1 standard to ensure the security of the UAV system with resource limitations, are considered in detail, on the basis of which a model in the class of models of formal grammars is built. The analysis of the model showed that in order to apply the method of protection based on the unidirectional function of grammatical analysis of regular language sentences, it is necessary to exclude the considered procedure from the protocol stack of the Reference Model of Open Systems Interaction.

Method of fast exposure in Galua fields in cryptographic data protection systems

O.P. Markovskyi — 2024-04-01

The method of accelerated calculation of the exponent on Galois fields – the basic operation of a wide range of cryptographic data protection algorithms is proposed and investigated. Acceleration is achieved by reducing the execution time of the ascent to the square in Galois fields, which is based on the use of the properties of the polynomial square, Montgomery reduction and the use of precalculations that depend only on the Galois Field base polynomial. The mathematical substantiation of the offered method and numerical examples which illustrate its work are resulted.

It is theoretically and experimentally proved that the proposed method allows to almost twice accelerate the exponentiation on Galois fields in comparison with known methods.

Method for accelerated modular squaring of long length numbers for cryptographic application

O.P. Markovskyi — 2024-04-01

The article theoretically substantiates and proposes a method for accelerated modular squaring of long length numbers.

Acceleration modular square calculation is achieved through the combined using of abridged multiplication technology, alternating cycles of multiplicative operations and partial modular reduction, as well as organizing simultaneous processing of a group of bits using pre-calculations under reduction.

The methodology for constructing of precomputations table for group Montgomery reduction is described in detail, as well as the proposed procedure for quickly calculating a modular square using abridged multiplication technologies and simultaneous Montgomery reduction of several digits. The presentation is illustrated with a numerical example.

It is shown that the proposed method makes it possible to speed up the computational implementation of the modular squaring operation, which is important for cryptographic applications, by 5-6 times compared to known technologies.

Performance analysis of YOLOv8n object detection algorithm on RaspberryPi and NVidia Jetson Nano microcomputers

Y.V. Lukash — 2024-04-01

The scientific article describes approaches to the implementation of a complex of autonomous control of UAVs. In particular, the architecture with the use of a single-board computer that can be installed directly on the aircraft is highlighted. Potential single-board computers were selected for such an architecture of the complex construction and their performance was investigated. For evaluation, an experiment was described and conducted to run the pattern recognition algorithm based on the Yolo v8 nano neural network. The results of the algorithm are given in the article. Based on the conducted experiment, it was determined that it is advisable to use Jetson Nano for similar algorithms. And RaspberryPi does not have enough power for a specific task with the selected algorithm.

Information and measurement system for diagnostic parameters of electric motors under conditions of destabilizing factors

V.P. Kvasnikov — 2024-04-01

The article is devoted to the development of an information and measurement system for detecting anomalies in the operation of electric motors, which integrates the methods of measuring the rotational moment and shaft speed. The main focus of the work is on the use of a neural network of the Autoencoder type, which is used to identify anomalous signals in noise conditions and under different loads on the electric motor shaft.

Experimental studies using this approach are presented, demonstrating its ability to detect deviations caused by sudden changes in load, as well as to identify anomalies in different modes of electric motor operation. Such a system contributes to increasing the reliability of electric motors and opens the way to optimizing the processes of their operation.

The methodology of measuring the torque, which was used when obtaining the characteristics at the time of acceleration of the electric motor, is presented.

The research results can be used to develop diagnostic equipment for electric motors.

Electronic voting protocols based on blind signature cryptographic scheme and biometric authentication

O.M. Zudov — 2024-04-01

The paper analyzes the requirements for the electronic voting procedure and identifies some problems that arise during the implementation of election schemes. Despite the large number of existing electronic voting schemes that use various cryptographic methods, there are still shortcomings and vulnerabilities that require special attention. One such problem is the prevention of vote selling. One of the most advanced voting methods, namely the He-Su protocol, was chosen as the basis, as it meets almost all the requirements for the electronic election process. The paper proposes to use cryptographic methods in combination with biometric authentication to partially solve the problem of vote selling. The proposed election algorithm is based on the He-Su protocol, which uses the homomorphism of the RSA cryptographic system and the blind signature procedure. Introducing the biometric authentication procedure into the algorithm reduces, in our opinion, the probability of votes being sold by unscrupulous voters, as it becomes difficult to do this without physical presence. Fingerprint biometrics with the use of a mobile application, and a biometric login authorized by the election organizer, appear to be the most convenient.

Using a spline model in the space of latent representations when removing duplicates from a set of observations

V.D. Zivakin — 2024-04-01

In the paper, for the first time, the use of a spline model based on local B-splines of the second order is proposed to estimate the density of the distribution of a training set of digital images in the latent space of a multilayer nonlinear auto encoder. Based on the model, a method for removing duplicate images in the latent space of the autoencoder network representation is proposed. Research has been conducted and statistically proven to increase the entropy of data distributions, which contributes to less retraining of neural models. The research focuses on learning digital images using a multi-layer nonlinear autoencoder, a deep learning tool that allows for dimensionality reduction and extraction of useful information from input data. The developed spline model provides new opportunities for estimating and visualizing distributions, which may be useful for further analytical research in the field of image processing.

The main focus of the work is concentrated on the method of removing duplicate images in the latent space, which uses data on the density of distributions obtained from the spline model. This allows not only to clean the data set from repeated samples, but also to optimize the learning process of neural networks, reducing overtraining and increasing the overall efficiency of the models.

Modeling and algorithmization of the pricing process in electronic commerce using software agents

E.V. Zelenko — 2024-04-01

The article is devoted to the improvement of the object-oriented programming method for the development of software used in electronic commerce, using an agent-oriented approach.

Experimental studies were conducted to compare traditional and improved methods for developing software used in e-commerce. In the course of research due to the application of an agent-oriented approach and a platform for the development of software agents, the following were found: an increase in the speed of software implementation of the user interface and asynchronous communication between autonomous system components; reducing the number of system processes used; reduction of connectivity between system components; increasing the efficiency of the software implementation of the developed data processing algorithm in terms of calculation accuracy; expanding the set of properties and potential capabilities of the software tool.

Methods of optimizing network architecture for high-performance computing in cloud environments

S.V. Zhuravel — 2024-04-01

The article provides a review of various methods for optimizing network architecture in the context of high-performance computing in cloud environments. The research is based on an analysis of modern technologies, strategies, and solutions used to enhance computational performance in the cloud. Special attention is given to networking architecture issues, including its structure, protocols, services, and optimization to achieve maximum productivity. The information presented may be useful for scientists, engineers, and software developers working on the creation and improvement of cloud computing systems.

Hybrid cryptosystems for the protection of FANET network nodes: approaches to use

S.S. Hrushak — 2024-04-01

The article presents the results of research into the possibility of using hybrid cryptosystems to protect FANET sensor network nodes. A model of a hybrid cryptosystem for data encryption, generation of a shared secret key is proposed, and an authentication algorithm in the FANET network based on the TLS protocol and symmetric/asymmetric encryption methods is developed. AES-GCM, ECDSA algorithms were tested for encryption and decryption operations, time characteristics of operations were measured, the number of CPU cycles and energy consumption were determined.

As a result of the conducted research, the expediency of using hybrid cryptosystems by nodes of FANET networks is substantiated, provided that certain requirements for hardware, software and energy resources are met.

Analysis of some well-known models of information spread in social networks in the aspect of information warfare

A.I. Gizun — 2024-04-01

Social networks have become an almost indispensable source of information and communication in the daily life of millions of people. From the birth of the concept in the 1950s, they have come a long way from defining certain social groups by interests to modern Internet platforms, with the prospect of further development to meta-platforms. With its development, social networks have become the main platform for the distribution of various information and one of the main factors of influence on society during election processes, revolutionary events and military conflicts. In order to conduct an information campaign or defend oneself in social networks, it is necessary to understand the principles of information dissemination in them.

On the basis of the analysis of models of information distribution in social networks, the most key ones have been selected. The following models of information dissemination in social networks are considered in the study: Model of viral distribution (SIR), SEIR model; Independent Cascade model; Wide-Spread Model; Influential Users Model. The results of the analysis made it possible to assess the suitability of these models for solving the problems of managing informational and psychological influence.

System of controlling the calibration of the coordinate measuring machine

O.V. Chalyi — 2023-12-25

The scientific article solves important scientific and technological problems of increasing the efficiency of the use of coordinate measuring machines and ensuring measurement accuracy. In the course of the work, a mathematical model of the system was presented, which makes it possible to assess the accuracy of contact measurements of complex curved surfaces by coordinate measuring machines and to determine evaluations of measurement results based on uncertainty. It has been shown that the actual measurement error can be much larger than the certified measurement error due to uncertainties.

The concept of building a hardware and software complex of UAV detection using artificial intelligence

Y.V. Utkin — 2023-12-25

The current state of affairs dictates the development of various developments and innovations in the field of information technology, which can have a dual purpose and be used both in peacetime and in wartime. The short list of projects in the direction of the rearmament of the armed of the Armed Forces of Ukraine testifies to intensive activity in the search for realistic, simple and economic solutions that will enable the Defense Forces of Ukraine to win the confrontation in a short period of time. Along with the positive reviews of foreign military experts regarding the development and application of technologies aimed at the defense capability of the state, there are a number of practical directions that require urgent solutions. One of these directions is the fight against the enemy's UAVs, which he actively uses to solve various tasks. The paper proposes the concept of building a UAV detection hardware and software complex using artificial intelligence in general and considers the possibility of further development of the proposed concept.

Forecasting prices of financial instruments using deep learning methods

M.L. Sattarova — 2023-12-25

This study is devoted to the problem of forecasting financial instruments’ prices on the stock market, the relevance of which has increased significantly in the modern world, and which is an important component of the processes of conducting financial activities and making informed investment decisions. A review and comparative analysis of the methods proposed in the existing studies was conducted, their shortcomings and shortcomings were revealed. Based on this, a new approach to solving this problem was proposed. The proposed approach is based on taking into account a complex set of factors for forecasting, including technical indicators, fundamental analysis data and macroeconomic factors, the use of a systematic approach to the selection of predictors (factors) for forecasting and inclusion in the model, the introduction of modern feature engineering and feature selection techniques, noise removal in input data, applying NLP techniques and sentiment analysis to integrate textual data that influence market trends, thus increasing the accuracy of modeling market processes. These principles were incorporated with machine and deep learning techniques capable of taking into account time sequences of data and complex relationships and dependencies between them, and a predictive model was built for forecasting stock prices. The results of model testing and evaluation of obtained performance accuracy metrics’ values for the developed model show it’s higher accuracy in comparison with the base model chosen for comparison, and also prove the effectiveness of using the proposed approach.

Method for cryptographically strict identification of remote abonents based on pseudorandom sequences generators

O.V. Rusanova — 2023-12-25

The article proposes a method for cryptographically strong remote abonents identification, which is based on the irreversibility properties of cryptographic generators of pseudo-random sequences. This made it possible to integrate, within a single technological solution, a cryptographically strong user identification before a session, permanent mutual authentication during a session, and also provide the possibility of streaming encryption of data exchange between the user and the system. It is shown that the use of the proposed method allows to reduce the identification time and increase the security against middle attacks.

Method of modular exponent secure distributed computation for accelerated implementation of data protection mechanisms in IoT

O.V. Rusanova — 2023-12-25

The article proposes a method of secure modular exponentiation on IoT terminal platforms involving cloud computing for fast implementation of public key cryptographic protection algorithms. The method implements a new principle of distribution of calculations between the terminal and remote platforms when implementing modular exponentiation - at the level of processing each individual bit of the exponent code. This makes it possible to reduce the computing load on the terminal platform by three times and, accordingly, to speed up the calculation of the basic operation of public key cryptography by three times compared to the known methods of distributed calculation of the modular exponent.

The method of monitoring forest ecosystem state using drones to detect fire based on video data

V.V. Nesterenko — 2023-12-25

Drones, as innovative tools, have found widespread applications across various sectors, including media, the military, agriculture, security, and scientific research. However, considering the escalating environmental issues, there arises a necessity for new technological approaches.

The current scientific trend focuses on developing a monitoring system for forest ecosystems using drones to respond to potential wildfires. This initiative involves deploying drones for aerial reconnaissance over forested areas and analyzing the captured video data to identify potential sources of ignition.

This endeavor represents a proactive use of drones in environmental monitoring, enhancing our ability to detect and respond to potential forest fires.

Recognition of fake news using natural language processing and a low-power architecture for edge computing

L.D. Mishchenko — 2023-12-25

The research focuses on enhancing the method of fake news detection based on natural language processing (NLP) utilising low-power edge computing architecture through attention mechanisms and contextual analysis. An attention mechanism and contextual analysis are implemented to detect linguistic signs of credibility and stylistic differences between fake and real news. This approach aims to enable news verification on peripheral devices with limited computational resources without compromising speed.

Experimental studies validate the efficiency of the proposed method in identifying anomalies in text frequency and credibility markers. Fake news is identified by the higher use of emotionally charged words, probative statements and profanity compared to real news.

The integration of attention mechanisms and contextual analysis showcases a notable improvement in identifying linguistic anomalies typical of fake news, achieving a classification accuracy of 81%. The findings contribute to combating misinformation by leveraging linguistic nuances and signify potential advancements in news veracity assessment on resource-constrained devices.

Method of multiprocessor modular multiplication with using Montgomery group reduction

O.P. Markovskyi — 2023-12-25

The article proposes a method of accelerating the operation of modular multiplication important for cryptographic applications due to its implementation in the form of m autonomous computational processes. The technology of allocation the multiplier into components and the procedure for performing partial modular multiplications using Montgomery group reduction based on precalculations have been developed in detail. The statement is illustrated by a numerical example. The choice of parameters of the organization of calculations that provides the greatest effect of parallelization of modular multiplication is theoretically substantiated. It is proved that due to the parallelization of multiplications and the group reduction of Montgomery, the calculation of the modular product is accelerated by 0.75×m-0.93×m times.