Information and telecommunication sciences: international research journal, Vol. 14, N. 1
Постійне посилання зібрання
Переглянути
Нові надходження
Документ Відкритий доступ Implementation of technology for improving the quality of segmentation of medical images by software adjustment of convolutional neural network hyperparameters(National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute", 2023) Prochukhan, Dmytro V.Background. The scientists have built effective convolutional neural networks in their research, but the issue of optimal setting of the hyperparameters of these neural networks remains insufficiently researched. Hyperparameters affect model selection. They have the greatest impact on the number and size of hidden layers. Effective selection of hyperparameters improves the speed and quality of the learning algorithm. It is also necessary to pay attention to the fact that the hyperparameters of the convolutional neural network are interconnected. That is why it is very difficult to manually select the effective values of hyperparameters, which will ensure the maximum efficiency of the convolutional neural network. It is necessary to automate the process of selecting hyperparameters, to implement a software mechanism for setting hyperparameters of a convolutional neural network. The author has successfully implemented the specified task. Objective. The purpose of the paper is to develop a technology for selecting hyperparameters of a convolutional neural network to improve the quality of segmentation of medical images. Methods. Selection of a convolutional neural network model that will enable effective segmentation of medical images, modification of the Keras Tuner library by developing an additional function, use of convolutional neural network optimization methods and hyperparameters, compilation of the constructed model and its settings, selection of the model with the best hyperparameters. Results. A comparative analysis of U-Net and FCN-32 convolutional neural networks was carried out. U-Net was selected as the tuning network due to its higher quality and accuracy of image segmentation. Modified the Keras Tuner library by developing an additional function for tuning hyperparameters. To optimize hyperparameters, the use of the Hyperband method is justified. The optimal number of epochs was selected - 20. In the process of setting hyperparameters, the best model with an accuracy index of 0.9665 was selected. The hyperparameter start_neurons is set to 80, the hyperparameter net_depth is 5, the activation function is Mish, the hyperparameter dropout is set to False, and the hyperparameter bn_after_act is set to True. Conclusions. The convolutional neural network U-Net, which is configured with the specified parameters, has a significant potential in solving the problems of segmentation of medical images. The prospect of further research is the use of a modified network for the diagnosis of symptoms of the coronavirus disease COVID-19, pneumonia, cancer and other complex medical diseases.Документ Відкритий доступ Source identification methodology in radio monitoring objects using multi-meaning(National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute", 2023) Ilnytskyi, Anatoliy I.; Tsukanov, Oleg F.Background. With the constant development of radio-electronic equipment and telecommunication technologies and the appearance of new technical means of various purposes, including dual meaning, which already have and operate with new “non-traditional” information features. The specified circumstances require the improvement of their radio monitoring and countermeasure systems, from the point of view of the development of highly effective intelligent systems that ensure the collection, processing, accumulation and use of the received monitoring information for decision-making on a real-time scale. Objective. Increasing the effectiveness of radio monitoring of radio radiation sources is carried out by decomposing their informational features into static and dynamic ones with further formalization of the observation and decision-making process. Methods. The decision on whether the source of radio radiation belongs to one or another class is made on the basis of a preliminary calculation of information feature estimates for all possible sets of features and the use of multi-valued logic functions to make a decision. The preliminary calculation makes it possible to increase the speed of the algorithm and make a decision about whether the object of recognition belongs to one or another class by calculating the value of only one function, and recalculating the estimates of static and dynamic information features only when the descriptions of classes of radio radiation sources are changed. Results. The proposed technique makes it possible to significantly expand the classes of classes of sources and objects of radio monitoring and will ensure an increase in the speed and reliability and efficiency of the recognition process as a whole. Conclusions. Deciding whether a source of radio radiation belongs to one or another class through the use of multi-valued logic functions allows increasing the efficiency of the radio monitoring system and significantly expand the class of monitoring sources.Документ Відкритий доступ Building minimum spanning trees by limited number of nodes over triangulated set of initial nodes(National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute", 2023) Romanuke, Vadim V.Background. The common purpose of modeling and using minimum spanning trees is to ensure efficient coverage. In many tasks of designing efficient telecommunication networks, the number of network nodes is usually limited. In terms of rational allocation, there are more possible locations than factually active tools to be settled to those locations. Objective. Given an initial set of planar nodes, the problem is to build a minimum spanning tree connecting a given number of the nodes, which can be less than the cardinality of the initial set. The root node is primarily assigned, but it can be changed if needed. Methods. To obtain a set of edges, a Delaunay triangulation is performed over the initial set of nodes. Distances between every pair of the nodes in respective edges are calculated. These distances being the lengths of the respective edges are used as graph weights, and a minimum spanning tree is built over this graph. Results. The problem always has a solution if the desired number of nodes (the number of available recipient nodes) is equal to the number of initially given nodes. If the desired number is lesser, the maximal edge length is found and the edges of the maximal length are excluded while the number of minimum spanning tree nodes is greater than the desired number of nodes. Conclusions. To build a minimum spanning tree by a limited number of nodes, it is suggested using the Delaunay triangulation and an iterative procedure in order to meet the desired number of nodes. Planar nodes of an initial set are triangulated, whereupon the edge lengths are used as weights of a graph. The iterations to reduce nodes are done only if there are redundant nodes. When failed, the root node must be changed before the desired number of nodes is changed.Документ Відкритий доступ Estimation of potential parameters for 5G mobile networks radiochannels(National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute", 2023) Berezovsky, Igor M.; Noskov, Vyacheslav I.Background. Deploying of 5G mobile networks opens up wide opportunities for the development of IoT, high-speed access to Internet services, industrial automation, telemedicine and other modern services. Peak transmission rate, latency, and spectral efficiency are important indicators for network performance. These indicators are primarily determined by the 5G-NR radio subsystem, which is built using modern technologies such as OFDM, interference resistant LDPC coding and massive MIMO antenna systems. In addition, frames and time-frequency resource distribution in 5G-NR are improved for both Downlink and Uplink. All of these are described in various 3GPPP documents, but to evaluate these indicators, it is necessary to create an appropriate methodology and perform calculations. Objective. The purpose of the research is to create a methodology and estimate the potential values of peak transmission rate, latency and spectral efficiency of 5G-NR radio channels. Method. Analytical calculation methods based on recommendations and source data of 3GPP documents are used. Results. Analytical studies show that 5G-NR radio channels can potentially provide a peak transmission up to 37 Gbps, latency less than 0.5ms, and spectral efficiency up to 46 bps/Hz rate in the Downlink direction using 50 MHz FR1 frequency band, QAM256 modulation and MIMO 8 x 8-antenna system. Conclusions. The researched 5G-NR radio channels efficiency indicators meet current and future services requirements.Документ Відкритий доступ The program for assessing the connectivity of nodes of wireless episodic networks under the condition of using UAVS(National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute", 2023) Sushyn, Ihor O.; Lysenko, Olexandr I.; Valuiskyi, Stanislav V.Background. Based on analytical mathematical models, the duration of connectivity of mobile subscribers (nodes) (MS) of a wireless episodic network (WEN, consisting of MS and UAV) was investigated in conditions of direct radio visibility and considering the relaying. Objective. The purpose of the work is to find methodological approaches to ensure the connectivity of WSN nodes, which is a necessary condition for obtaining information from WSN in the absence of communication infrastructure. Methods. Simulation modelling based on MAPLE 14 software package and analytical calculation methods are used. Results. It is shown that the duration of connectivity is directly proportional to the size of the coverage area and inversely proportional to the movement speed of nodes. The mobility nature (scenario) of nodes also affects the duration of connectivity. The simulation of the nodes' movement was carried out under 4 scenarios: "march", "incoherent", "random wandering in the field" and "random wandering in the city". The largest values of the connectivity duration correspond to the third scenario, and the smallest - to the second (with a fixed radius of the coverage area and the movement speed of nodes). Thus, the average connectivity duration of the UAV-pedestrian connection in the event of an "incoherent" will be about 36 minutes, and the UAV-car connection - about 5 minutes. Conclusions. The system and functional parameters of the networks, which were obtained as a result of the research, will form the basis of the initial data and limitations of the mathematical model, and will also make it possible to determine the initial placement of the UAV network at the planning stage.Документ Відкритий доступ Comparison of optimization strategies and estimation techniques for radio network planning and optimization problems(National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute", 2023) Prokopets, Volodymyr A.; Globa, Larysa S.Background. Radio network planning is one of the main phases of the cellular network lifecycle, as it determines capital and operating costs and allows system performance evaluation at any given time. An accurate and comprehensive analysis of existing network statistics is necessary for proper cell planning during network expansion. These statistics are collected throughout the life cycle of the cellular network and usually have certain imperfections (heterogeneity of statistics, which have different densities in different parts of the search space, up to the presence of significant voids, etc.) The system describing the functioning of the radio network can be represented as a black box because its internal processes are too complex to be defined by mathematical functions. This determines the need to use appropriate tools. Objective. The purpose of the paper is to create a toolkit that allows finding the proper relationships between network parameters to define target values that will help to build an effective network plan in terms of performance and costs for its creation and operation. The tools should be able to work efficiently using the minimum set of available statistical data, as well as taking into account their imperfections. Methods. Mathematical estimation and optimization methods are used, namely Ordinary Least Squares, Ridge Regression, Lasso, Elastic-net, LARS lasso, Bayesian Ridge Regression, Automatic Relevance Determination, Stochastic gradient descent, Theil-Sen estimator, Huber Regression, Quantile regression, Polynomial regression. We consider 12 estimation methods in combination with two optimization strategies. Additionally, the method of partial analysis of the search space with different number of configurations is considered. Results. A software package using the Python programming language has been created, which contains a practical implementation of all the considered estimation and optimization methods, as well as tools for evaluating arbitrary configurations of the software package (benchmark) and visualizing the results. The best estimation method is Ordinary Least Squares for finding the optimal configuration of the statistical parameters of the 4G radio network to maximize the download speed. To obtain satisfactory results, it is enough to consider 25 initial and 250 estimated points - a larger number of points will not significantly increase prediction accuracy. Conclusions. The results indicate the possibility of using the created software package for radio network planning tasks. Further research is aimed at expanding the created software package's functionality and considering additional estimation methods and optimization strategies.Документ Відкритий доступ Internet of things data transfer method using neural network autoencoder(National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute", 2023) Siemens, Eduard; Kurdecha, Vasyl V.; Ushakov, Serhii M.Background. The number of devices in the Internet of Things is constantly increasing. At the same time, the number of solutions on the market for such technologies is growing. Statistics confirm that these factors lead to an increase in data transfer volumes. This raises the number of resources spent on data transmission. The growing trend in the number of users of the Internet of Things technology leads to the emergence of the problem of a rapid increase in the data transmitted by the network. Objective. The purpose of the paper is to improve the process of data transmission in the Internet of Things by modifying the neural network autoencoder to reduce network resources use. Methods. Analysis of publications dedicated to Internet of things data transmission. Integration of existing data coding solutions based on a neural network autoencoder in the process of transmitting data from the Internet of things. Results. The neural network autoencoder has been improved by using an algorithm that additionally includes an arithmetic encoder and further training a new model on the output of a full-fledged autoencoder. Conclusions. The process of data transmission in the Internet of Things network has been modified by improving the neural network autoencoder by using the training of a smaller neural network on the initial data of the main autoencoder, which has reduced the amount of data transmitted and, accordingly, reduced the use of network resources.Документ Відкритий доступ Continuation of academician V.M. Glushkov work on the development of modern information & communication technologies(National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute", 2023) Ilchenko, Мykhaylo Yu.; Kravchuk, Serhiy O.; Uryvsky, Leonid O.Background. The distinguished contribution of Academician V.M. Glushkov to the theory and practice of automated control systems’ national-wide creation makes his achievements as the scientist and science organizer to be relevant and important. The world’s scientific community celebrates his 100th anniversary this year. Results. The significant achievements, prospective developments, recognized scientific schools of Igor Sikorsky Kyiv Polytechnic Institute in the most important areas of fundamental and applied scientific research in information technologies were noted. The leader of the current and advanced developments and achievements implementation in the information and telecommunication technology area in our university is ER ITS. That indicates the succession of generations, the proven foundation of the brilliant ideas’ implementation, that are bequeathed to us by Academician V.M. Glushkov.