The paper considers the problem of improving
the recognition of not-perfect labeled face images using CNN
networks. The proposed solution is based on relabeling the
samples of images by applying the KNN classification principle
based on the distance between the samples. The original images
are first converted to the features and the KNN principle is
applied to them. The classes of...
This research presents an application of the Mask R-CNN algorithm for apple detection and semantic segmentation, aiming to enhance automation in the agricultural sector. Despite the growing use of deep learning techniques in object detection tasks, their application in agricultural contexts, specifically for fruit detection and semantic segmentation, remains relatively unexplored. This study...
Artifacts that occur during the registration of electroencephalographic signals (EEG) pose a significant problem, hindering the analysis of these signals for medical diagnosis or brain-computer interfaces (BCIs). While some artifacts can be relatively easily removed, others, such as those related to muscle activity, are more challenging to eliminate. Continuous research is being conducted to...
High voltage power lines inspection requires high amount of effort and time and needs to be performed regularly. As high voltage power lines are critical infrastructure reducing the frequency of the inspections is not an option. To reduce effort required, some tasks can be automated. For example assessment of the state of transmission towers and power lines or autonomous UAV flights to gather...
Artifacts pose a significant challenge in the analysis of EEG signals. Physiological artifacts stem from natural activities of the human body, such as swallowing saliva, clenching the jaw, facial grimacing, and eye blinking, among others. Visual evaluation often serves as the basis for artifact elimination. In this study, the authors investigated the impact of artifacts on the detection of...
This study investigates the impact of various signal features on machine learning-based tool condition classification in the milling chipboard process. Different machine learning models such as XGBoost, Gradient Boosting, Decision Tree and Random Forest have been applied and the signal features have been ranked based on their importance. The highest ranking signal was 'DataLow_0', contributing...
This paper presents a comprehensive performance evaluation of various AI architectures for a classification of holes drilled in melamine faced chipboard, including custom Convolutional Neural Network (CNN-designed), five-fold CNN-designed, VGG19, single and five-fold VGG16, an ensemble of CNN-designed, VGG19, and 5xVGG16, and Vision Transformers (ViT). Each model's performance was measured and...
Large Language Models (LLMs) have been gaining tremendous popularity since early 2023 with the release of the GPT-4 system. The new artificial intelligence methods have successes in many fields of activity, including education. This article analyses the applicability of LLMs in supporting the teaching of the "Electromagnetic Fields" course in Electrical Engineering studies. Examples of use in...
Abstract. — The AdaDelta learning process optimization method has been tested for a multilayer neural network with three hidden layers with 28 neurons each, when recognizing printed numbers. Testing the learning error of this neural network was carried out using the mapping function and Fourier spectra of the error function. The mapping function describes the process of doubling the number of...
Specialized game computer system based on Arduino Uno - indie project was proposed and investigated. Based on the analysis of the current state of the gaming industry and global trends, the main theses of a successful product, its implementation, operation and the main stages of development were formed. The solution of key problems at all stages of development was described, the corresponding...
This paper presents the design and development of an intelligent air quality monitoring system that utilizes the widely adopted and versatile Arduino Uno microcontroller as its foundational platform. The system underwent comprehensive testing procedures to ensure its adherence to specified requirements. Moreover, a series of experiments were conducted in diverse areas of a residential...
This article presents numerical simulations focused on static magnetic fields, employing a Mu-metal cage for shielding against the geomagnetic field. The objective is to verify, how a homogeneous and uniform magnetic field within the magnetic field applicator is created. Square Helmholtz coils are employed to ensure this homogeneity. Observations indicate a 5.3% deviation in the B-field when...
The paper presents and discusses an enhanced
version of modeling laser or small-scale induction hardening.
The process is characterized by a rapid cooling rate, reaching
several hundred or over one thousand degrees Celsius per second.
Most heat from the heated spot is transferred through conduction
rather than convection, penetrating deep into the material’s
interior. Consequently,...
This study compares two SAR techniques (10g method and 1g method) on a chest model of an adult and a child in the shielded area of a train car compartment to examine the impact of electromagnetic (EM) fields. This study made use of electromagnetic modeling based on the Finite Integration method. The simulations were done for a particular scenario in which each model was positioned as near to...
Measuring the flow of carbonated drinks is not completely resolved. Carbonated drinks are liquids of an inho- mogeneous type. Individual drinks differ in many parameters (e.g. conductivity, pH, color, etc.) and any change in temperature or pressure affects these parameters. The article describes the design and optimization of the magnetic circuit of an induction flow meter for measuring...
In order to narrow down the characteristics of hot springs that have efficacies on arteriosclerosis, changes in skin cholesterol levels after bathing at the Yobizuru Onsen, which is classified as a simple weakly radioactive cold mineral spring, were measured. As a result, based on the results from the mathematical model of arteriosclerosis onset, it was indirectly shown that the nature of a...
he advancement of the sensor technology becoming increasingly cost-effective and the progress in diagnostic and management research, users nowadays not only demand high reliability from their devices but also the ability for their equipment to self-diagnose errors and provide alerts. These devices often incorporate sensor systems capable of generating tens of thousands of data points per...
Abstract—The problems associated with the operation of overhead power lines and ways of improving control over their condition with the help of UAVs are considered. A structural diagram of the system of technical diagnostics of overhead lines based on UAVs was developed, for which the necessary diagnostic parameters were selected according to informative criteria. An analysis of types of UAVs...
The paper presents an improved model of induction hardening of gear wheels. Although the mathematical model of the hardening is well known, some parameters are subject to uncertainties (temperature dependence of material parameters, coefficient of convective heat transfer, or emissivity). The authors offer a methodology based on calibration and optimization techniques to minimize the relevant...
This paper deals with the design and development of an extruder for printing ceramic materials including its control. The paper describes the experimental development of the extruder, divided into a mechanically extruded stack and the method of feeding the material into a print head equipped with a screw conveyor. The paper includes a computer model of the designed extruder. The aim of the...
In the paper, we present the machine learning algorithm to predict 100m men’s outdoor sprint score. For this project, a unique dataset of 17 features was created. The training set contained 174,383 records, whereas the test set contained remaining 406,894 records (proportions: 30% of training data to 70% of test data). The proposed multi-layer MLP model is based on set of features, such as...
The paper presents a study concerning recognizing user emotion based on keystroke dynamics of the written text. At first, the analysis of the dataset used in the task is performed. Followed by the training and the effectiveness assessment of classical methods: Naive Bayes, K-Nearest Neighbours, Random Forest, and Multilayer Perceptron applied to the classification of provided samples to one of...
The work deals with the analysis of a digital recording of cough sounds. This signal was obtained using a digital recorder with a microphone placed on the patient's chest. The wavelet transform was used in this work. The principle of its operation, various wavelet functions, and types of these functions are described. The core of the work is the continuous wavelet transform algorithm. The...
The article presents a system for monitoring fluid during intravenous infusion. The system is based on a capacitive method in which the change in electrical capacitance of the infusion container is converted into the current volume of the fluid. The capacitor pads of the measurement system are integrated into the housing of the container, facilitating easy installation in hospital conditions...
An electromechanical system (EMS) for adjusting the arc lengths of an arc furnace (AF) using a fuzzy model for the electrodes movement control is proposed. A fuzzy impedance-admittance law to form the electrical mode (EM) mismatch signal is created and the use of a fuzzy corrector for the electrode motion control signal is validated. A parametric synthesis of the fuzzy corrector based on the...
This paper presents the application of ANSYS circuit simulator for the identification and evaluation of the potential sources of electromagnetic disturbances in the high- current laboratory environment. Use of the computer model of the test circuit and comparison of the simulation results with the measurements allowed to confirm the hypothesis that the characteristic patterns of transient...
The research aimed to verify the adopted methodology for testing transients in circuits with inductive elements during their disconnection. The main focus of interest is the electromagnetic disturbances named electrical fast transients (EFT/Burst) that occur on this occasion. The analytical derivation allowed us to know the nature of the transition state and various cases that depend on the...
Electromagnetic nondestructive evaluation of conductive biomaterials using the sweep frequency eddy current method is the subject of this work. The main aim is to verify its use in the investigation of artificial defects that are located inside the austenitic stainless steel. For this purpose, numerical simulations and experiments were carried out, the results of which will be compared and evaluated.
This paper introduces a Finite-Difference Time Do- main (FDTD) electromagnetic field solver focused on speed and low-memory usage. It solves both electric and magnetic fields on axisymmetric grids. The accuracy of the code has been compared to the equivalent Finite Element Method (FEM) models in 2D and 3D. Finally, the performance metric of the FDTD solver has been estimated and compared with...
The article presents the proprietary development version of the system for monitoring and detecting a low level of infusion fluid, initiated by a medical university. The results presented earlier by one of the coauthor are the basis for expanding the functionality of the measurement device. Additional features include the handling of multiple containers with liquid by a single station (just as...
The article presents analysis of static compensator influence on average power flow. It is shown that the compensator current affects on the power delivered to the consumers. The real time iterative method of compensator control in was presented. The RMS value of grid voltage is chosen as the target of the compensator. The results prove the effectiveness of the proposed control method.
This scientific article explores the impact of collision chemistry on Townsend’s avalanche development through an innovative computational approach. By combining the Particle- in-Cell (PIC) algorithm and Monte Carlo (MCC) collisions, we determine the first Townsend coefficient for Helium gas. Further, we investigate the influence of a variable number of collisions in our simulations to achieve...
This scientific research is devoted to the analysis and testing of the Haar cascade classifier method for face recognition. The paper also justifies the choice of algorithm and provides a description of the implementation of the system on which the analysis was conducted. During the research, the developed system was tested on a sample of facial photographs that included images at different...
The article describes the concept of steganography from the point of view of the issues of the physical and digital world. It discusses the basics of hiding information in data medium, indicates the potential in terms of their capacity and susceptibility to the phenomenon of data concealment. It also indicates aspects of the suitability of modified data for information retrieval and...
Digital substations, based on technologies like IEC 61850 and IEC 60870, are at the forefront of modern power grid advancements, offering enhanced automation, communication, and control capabilities. Mongolia's Ministry of Energy initiated efforts to transition to digital substations as part of their "Smart Energy" project. However, implementing these advanced systems in Mongolia is not...
This article describes the construction of a test stand for locating disturbance sources using an EMC scanner placed in a dedicated EMC test chamber. In the description of the construction of the stand, special attention was paid to the EMC test chamber and the solutions used for shielding and minimizing reflections of electromagnetic waves inside the chamber. The research results presented in...
In the scientific work, a module is proposed that allows to detect incorrect and irrelevant information based on the content analysis of web resources in the intellectualized academic staff performance appraisal system. The methods and means of detecting irrelevant information based on existing user needs are analyzed. This module allows to analyze the information filled in by the user and...
This following article presents the concept of a cheap universal capacitive sensor for partial discharge (PD) measurement. The authors checked the effectiveness of the device prototype on a MV air-insulated switchgear. For this purpose, a combination of a widely used Transient Earth Voltage (TEV) sensor and a detector with Junction Field-Effect Transistor (JFET) was used. The paper presents...
Abstract— the frequency symbolic method (FS method) of steady-state analysis of linear periodically time-variable (LPTV) circuits is used to form their transfer functions in the frequency domain. A Fourier polynomial approximates the transfer function and contains the complex variable, the time variable, and the parameters of the circle elements in the form of symbols. The coefficients of such...
The article presents an analysis of the possibility of taking into account the correction factor for the measured values of the current pulse. For this purpose, the article presents the results of tests using various types of Pearson coils. The features of such a correction factor, method, and scope of application are defined. The obtained results are the basis for further research in order to...
The paper considers the stochastic optimization method, which can be used for dynamic processes modeling in complex electrical systems, which are described by input- output models. As the main method, Rastrygin's guide cone method is considered. Limitations for this method are formulated for the practical usage when solving applied optimization problems. In order to improve the method,...
