Repository of FERIT Osijek
Not a member yet
4053 research outputs found
Sort by
Estimation of the propagation of measurement uncertainty using the adaptive Monte Carlo method when determining the remanent current based on the steady state established after switching on the transformer to the rated alternating voltage
No full textOvaj rad definira pojam remanentnog magnetskog toka i mjerne nesigurnosti.
Nadalje, pruža opis adaptivne Monte Carlo metode i metode određivanja remanentnog toka u
magnetskoj jezgri na temelju ustaljenog stanja uspostavljenog nakon uklopa transformatora na nazivni
izmjenični napon. Primjenom adaptivne Monte Carlo metode, procijenjena je propagacija relativne
mjerne nesigurnosti mjerene veličine za pomenutu metodu pomoću programskog paketa MATLAB.This paper defines the concept of remanent magnetic flux and measurement uncertainty.
Furthermore, it provides a description od the adaptive Monte Carlo method and the method of
determining the remanent flux in the magnetic core based on the steady state established after the
transformer is connected to the rated alternating voltage. Using the adaptive Monte Carlo method, the
propagation of relative measurement uncertainty of the measured quantity for the specified method
was estimated with MATLAB software
Development of a web environment for evaluating and comparing graphical user interfaces
No full textOvaj rad bavi se istraživanjem objektivnosti studenata pri međusobnom ocjenjivanju radova iz
područja dizajna grafičkih korisničkih sučelja. U tu svrhu kreirana je web aplikacija koja
omogućava kardinalno i ordinalno ocjenjivanje radova prijavljenim korisnicima. Proveden je
eksperiment s 15 ispitanika – studenata FERIT-a koji su ocjenjivali stvarne radove drugih
studenata. Ispitanici su morali ocijeniti radove ocjenama od 1 do 5, a potom i evaluirati dizajne
metodom potpune i nepotpune parne usporedbe. Rezultati parne usporedbe korišteni su za
stvaranje rang liste studentskih radova. Ispitana je korelacija ocjena studenata i nastavnika
koristeći Pearsonov koeficijent korelacije. Korelacija rangova studenata i nastavnika te korelacija
dvije metode parne usporedbe ispitane su koristeći Spearmanov i Kendallov koeficijent korelacije.
Promotrena je i povezanost između kardinalnog i ordinalnog ocjenjivanja te utjecaj smanjenja
lambda replikacije. Rezultati su pokazali da je parna usporedba prikladan način ocjenjivanja
vršnjaka te da je moguće smanjiti opterećenje na ocjenjivače metodom nepotpune usporedbe i
pritom zadržati visoku razinu korelacije s rezultatima potpune usporedbe.This paper focuses on researching the objectivity of students in evaluating each other's assignments
in the field of graphical user interface design. To conduct this research, a web application was
created that allowed registered users to evaluate assignments using both cardinal and ordinal
evaluation. The experiment involved 15 FERIT students who evaluated the real assignments of
other students. The respondents assessed the assignments using marks from 1 to 5 and then
evaluated the designs using complete and incomplete pairwise comparison methods. The results
of the pairwise comparison were then used to create a ranking list of student papers. The Pearson
correlation coefficient was used to determine the correlation between the grades given by students
and a teacher. The correlation between ranks determined by students and a teacher and the
correlation between the two methods of pairwise comparison were assessed using Spearman and
Kendall correlation coefficients. Further, the study observed the relationship between cardinal and
ordinal evaluation and the impact of reducing lambda replication. The study found that pairwise
comparison is a suitable approach to evaluating peers and that it is possible to reduce the burden
on evaluators using the method of incomplete comparison while maintaining a high level of
correlation with the results of the complete comparison
Automatic design of convolutional neural network using particle swarm optimization
No full textKonvolucijske neuronske mreže primjenjuju se za rješavanje raznih problema dubokog učenja. Primjena konvolucijskih neuronskih mreža za zadani problem zahtijeva dizajniranje odgovarajuće strukture mreže. Ručni odabir i vrednovanje potencijalnih struktura mreža zamoran je i dugotrajan proces. Automatizacija ovog postupka uvelike može smanjiti napore i vrijeme potrebno za dizajniranje prikladnih mreža. Automatski dizajn neuronskih mreža može se predstaviti kao optimizacijski problem te je u literaturi predloženo pregršt postupaka za njegovo rješavanje. U ovom radu korišten je algoritam PSO koji se pokazao učinkovitim u rješavanju brojnih problema optimizacije u literaturi, među kojima je i dizajn neuronskih mreža. Kako bi se performanse algoritma PSO mogle usporediti s nekom od poznatijih i jednostavnijih algoritama za dizajniranje neuronskih mreža, proveden je postupak pronalaska prikladne inačice algoritma PSO u smislu načina predstavljanja rješenja i odabira vrijednosti njegovih parametara. Eksperimentalna analiza provedena je na četiri skupa podataka: MNIST, Fashion MNIST, CIFAR-10 i CIFAR-100. Eksperimentalnom analizom utvrđeno je kako je najbolji način predstavljanja rješenja pomoću vektora realnih vrijednosti, a kod odabira parametara povoljnim se pokazalo odabrati jednake vrijednosti akceleracijskih koeficijenata. Eksperimentalnom analizom također je utvrđeno kako algoritam PSO postiže bolje rezultate od uobičajeno korištenih postupaka za dizajniranje neuronskih mreža, što ga čini prikladnim algoritmom za ovaj problem.Convolutional neural networks are applied to solve various deep learning problems. Applying convolutional neural networks to a given problem requires designing an appropriate network structure. Manual selection and evaluation of potential network structures is a tedious and time-consuming process. Automating this process can significantly reduce the effort and time needed to design suitable networks. The automatic design of neural networks can be presented as an optimization problem, and a handful of procedures for solving it have been proposed in the literature. In this paper, the PSO algorithm was used, which proved to be effective in solving numerous optimization problems in the literature, among which is the design of neural networks. To compare the performance of the PSO algorithm with more well-known and simpler algorithms for neural network design, a process of finding a suitable version of the PSO algorithm in terms of solution representation and parameter value selection was conducted. Experimental analysis was carried out on four datasets: MNIST, Fashion MNIST, CIFAR-10 and CIFAR-100. Through experimental analysis, it was determined that the best way to represent solutions is through a vector of real values and selecting equal values for acceleration coefficients proved favorable in parameter selection. The experimental analysis also revealed that the PSO algorithm achieves better results than commonly used procedures for neural network design, making it a suitable algorithm for this problem
Automated email service testing using the Cypress framework
No full textOvaj diplomski rad istražuje email tehnologiju i pristupe njezinom testiranju, s
posebnim naglaskom na cypress razvojno okruženje. Objašnjava koncept email tehnologije i metode za
njezino testiranje unutar cypressa. Centralni dio rada predstavlja novi priključak za testiranje email
servisa, razvijen s ciljem otklanjanja nedostataka postojećih rješenja i omogućavanja besplatnog
testiranja. Priključak je integriran u cypress dokumentaciju nakon pregleda od strane razvojnih inženjera
iz Cypress.io. Rad naglašava važnost automatizacije testiranja email servisa zbog njegove kompleksnosti i
učestalih potreba za raznim scenarijima testiranja. Predloženo rješenje koristi IMAP protokol za
interakciju s email serverom, omogućujući testiranje prema različitim email serverima uključujući
popularne poput Gmaila. Ovim radom se unapređuje razumijevanje i primjena procesa testiranja email
servisa, pružajući korisne alate za inženjere u razvoju softvera.This thesis explores email technology and approaches to its testing, focusing on the Cypress
development environment. It explains the concept of email and methods for testing it within cypress.
The central part of the thesis introduces a new plugin for testing email services, developed to overcome
shortcomings of existing solutions and enable free testing. The plugin is integrated into cypress
documentation after review by Cypress.io. The thesis emphasizes the importance of automating email
service testing due to its complexity and frequent need for various testing scenarios. The proposed
solution uses the IMAP protocol for interacting with email servers, allowing testing against different
servers including popular ones like Gmail. This thesis enhances understanding and application of email
service testing processes, providing valuable tools for software development engineers
Application of OpenAI Models in the Process of Engineering Requirements for Web Applications
No full textOvaj rad se bavi istraživanjem uspješnosti generiranja kvalitetnih dokumenata
zahtjeva korištenjem velikih jezičnih modela. Cilj je bio procijeniti može li se proces inženjeringa zahtjeva
olakšati implementacijom OpenAI modela u kreiranju web aplikacija. Kroz razvijenu web aplikaciju koja
koristi OpenAI asistenta, omogućeno je generiranje i uređivanje dokumenata zahtjeva na brz i
jednostavan način. Evaluacija je provedena usporedbom dokumenata generiranih različitim OpenAI
modelima s dokumentom iskusnog inženjera zahtjeva. Rezultati pokazuju da se aplikacija može koristiti
za inženjering zahtjeva, uz svjesnost korisnika o ograničenjima modela. Ova aplikacija može olakšati
proces kreiranja softvera osobama s manje iskustva, ali i ubrzati rad iskusnih inženjera.This paper investigates the effectiveness of generating quality requirement documents using
large language models. The goal was to evaluate whether the requirements engineering process can be
made easier by implementing OpenAI models in creating web applications. Through a developed web
application utilizing the OpenAI assistant, it enables quick and easy generation and editing of
requirement documents. Evaluation was conducted by comparing documents generated with different
OpenAI models with a document created by an experienced requirements engineer. The results show
that the application can be used for requirements engineering, with users being aware of the model's
limitations. This application can facilitate the software creation process for less experienced individuals
and speed up the work of experienced engineers
Photovoltaic systems with Sun tracking
No full textOvaj završni rad istražuje tehnološke aspekte i inovacije u području
fotonaponskih sustava za praćenje Sunca, koji su ključni za optimizaciju proizvodnje električne energije iz
sunčevog zračenja. Rad započinje pregledom literature i uvodom u osnovne pojmove sunčevog zračenja
i fotonaponskih elektrana. Detaljno se analiziraju različite vrste sustava za praćenje Sunca, uključujući
jednoosne i dvoosne sustave, te njihova primjena u različitim geografskim i klimatskim uvjetima.
Fotonaponski sustavi za praćenje Sunca omogućuju optimalno usmjeravanje fotonaponskih modula
prema položaju Sunca tijekom dana, što značajno povećava njihovu učinkovitost. Jednoosni sustavi
praćenja prilagođeni su za dnevnu optimizaciju izlaza energije, dok dvoosni sustavi pružaju dodatne
prednosti u regijama s većim varijacijama položaja Sunca tijekom godine. U radu se upotrebljava PV*SOL
premium program za simulaciju i analizu performansi različitih fotonaponskih sustava. Parametri
simulacije uključuju različite konfiguracije sustava praćenja Sunca (fiksni, jednoosni s osima sjever-jug i
istok-zapad, jednoosni s rotirajućom vertikalnom osi, te dvoosni sustavi) i njihov utjecaj na proizvodnju
električne energije. Analiza rezultata simulacija pokazuje značajna poboljšanja u proizvodnji električne
energije pri korištenju sustava za praćenje Sunca u usporedbi s fiksnim sustavima. Povećanje
energetskog prinosa varira između od 22% do 56%, ovisno o načinu upravljanja sustavom, stupnju
automatizacije, slobodi kretanja panela i meteorološkim uvjetima. Zaključuje se da daljnji razvoj
tehnologije praćenja Sunca igrati će ključnu ulogu u napredovanju obnovljivih izvora energije i postizanju
globalnih ciljeva održivog razvoja.This thesis explores technological aspects and innovations in the field of solar
tracking photovoltaic systems, which are essential for optimizing the production of electrical energy
from solar radiation. The work begins with a review of the literature and an introduction to the basic
concepts of solar radiation and photovoltaic power plants. Different types of solar tracking systems,
including single-axis and dual-axis systems, and their application in various geographical and climatic
conditions are thoroughly analyzed. Solar tracking photovoltaic systems enable optimal orientation of
photovoltaic modules towards the position of the sun throughout the day, significantly increasing their
efficiency. Single-axis tracking systems are adapted for daily optimization of energy output, while dualaxis
systems offer additional benefits in regions with greater variations in the sun's position throughout
the year. The thesis utilizes the PV*SOL premium program for simulation and analysis of the
performance of various photovoltaic systems. Simulation parameters include different configurations of
solar tracking systems (fixed, single-axis with north-south and east-west axes, single-axis with rotating
vertical axis, and dual-axis systems) and their impact on electricity production. Analysis of the simulation
results shows significant improvements in electrical energy production when using solar tracking
systems compared to fixed systems. The increase in energy yield varies between 22% and 56%,
depending on the management of the system, degree of automation, freedom of panel movement, and
meteorological conditions. It concludes that further development of solar tracking technology plays a
key role in advancing renewable energy sources and achieving global sustainable development goals
Block-Based Interface for Programming Infineon XMC Series Microcontrollers
No full textU ovom diplomskom radu istražena je mogućnost razvoja blokovskog sučelja za
programiranje Infineon XMC serije mikroupravljača, koristeći Blockly kao osnovu za vizualno
programiranje. Razvijen je sustav koji omogućuje korisnicima jednostavno kreiranje programa putem
grafičkog sučelja, čime se smanjuje složenost kodiranja i omogućuje šira primjena mikroupravljača u
industrijskim i edukacijskim okruženjima. Posebna pažnja posvećena je izradi vlastitih blokova i
prevoditelja za generiranje C koda, kao i integraciji s ModusToolbox razvojnim okruženjem. Testiranje
sustava provedeno je kroz nekoliko praktičnih primjera, pri čemu su identificirane mogućnosti za daljnje
unaprjeđenje i optimizaciju sustava.This thesis explores the development of a block-based interface for programming Infineon
XMC series microcontrollers using Blockly as the foundation for visual programming. A system was
developed to enable users to easily create programs through a graphical interface, thereby reducing the
complexity of coding and broadening the application of microcontrollers in industrial and educational
settings. Special attention was given to the creation of custom blocks and a translator for generating C
code, as well as the integration with the ModusToolbox development environment. The system was
tested through several practical examples, identifying opportunities for further enhancement and
optimization
Digital Kanban Board
No full textU ovom radu objašnjena je Kanban ploča kao alat za vizualno upravljanje zadacima i
procesima. Rad se fokusira na razvoj digitalne verzije Kanban ploče, koja omogućava korisnicima
učinkovito praćenje i organizaciju zadataka u timskom radu. Aplikacija pruža funkcionalnosti kao što su
dodavanje, premještanje zadataka i pregled gotovih zadataka, čime se poboljšava preglednost i
produktivnost. Razvoj digitalne Kanban ploče demonstrira kako tehnologija može unaprijediti
tradicionalne metode upravljanja projektima, prilagođavajući ih modernim potrebama korisnika.In this paper, the Kanban board is explained as a tool for visual management of tasks and
processes. The paper focuses on the development of a digital version of the Kanban board, which allows
users to efficiently track and organize tasks in teamwork. The application provides functionalities such as
adding, moving tasks, and reviewing completed tasks, whereby improving visibility and productivity. The
development of the digital Kanban board demonstrates how technology can enhance traditional project
management methods, adapting them to the modern needs of users
Creation of a 3D model on the example of a power line pole
No full textOvaj rad bavi se izradom 3D modela dalekovoda korištenjem softvera Blender, s
ciljem stvaranja realistične i detaljne vizualizacije. Glavni problem je stvaranje vjernog prikaza
dalekovoda koji odražava stvarne tehničke specifikacije. Rješenje je postignuto korištenjem stvarnih
tehničkih podataka tijekom modeliranja te renderiranjem na grafičkoj kartici AMD Radeon RX 7900XT,
što je omogućilo visoku kvalitetu prikaza. Rezultat rada su ključne perspektive modela: pogled iz ljudske
perspektive, pogled iz zraka, prednji pogled, detalj visećeg izolatora i puni pogled na oba dalekovoda u
prirodnom okruženju.This paper deals with the creation of a 3D model of a transmission line using Blender software,
with the aim of creating a realistic and detailed visualization. The main problem is to create a faithful
representation of the transmission line that reflects the actual technical specifications. The solution was
achieved by using real technical data during modeling and rendering on an AMD Radeon RX 7900XT
graphics card, which enabled high quality display. The result of the work are the key perspectives of the
model: a human perspective view, an aerial view, a front view, a detail of a suspended insulator and a
full view of both transmission lines in a natural environment
Virtual floating based camera keyboard
No full textU ovom diplomskom radu opisana je izrada virtualne tipkovnice koja koristi kameru
za prepoznavanje pritisaka prstiju na ravnoj radnoj površini. Sustav omogućuje tipkanje bez
fiziĉke tipkovnice, uz minimalnu potrebnu opremu i jednostavno odrţavanje. LRCN mreţa
prepoznaje pritiske lijeve i desne ruke te razlikuje situacije kada nema pritiska. Za potrebe
treniranja neuronske mreţe, kreiran je novi skup podataka s raznovrsnim slikama ruku jer
postojeći nisu odgovarali potrebama ovoga rada. Algoritam, razvijen u Python programskom
jeziku, u stvarnom vremenu obraĊuje videozapise, prikuplja petnaest uzastopnih okvira i
koristi neuronsku mreţu za prepoznavanje pritisaka tipki. Grafiĉki prikaz omogućuje
interaktivno tipkanje slova, znakova, brojeva i osnovnih funkcija kao što su enter, space i
backspace. Model je treniran i evaluiran na zasebnom testnom skupu, a najbolji model
odabran je na temelju performansi. Testiranje s tri skupine korisnika pokazalo je
funkcionalnost i pouzdanost tipkanja, uz nedostatak kao što je sporost tipkanja zbog
vremenskih zahtjeva modela i kamere.In this thesis the creation of a virtual keyboard is described, that uses a camera to
recognize finger presses on a flat working surface. The system enables typing without a
physical keyboard, requiring minimal equipment and simple maintenance. An LRCN network
recognizes presses from the left and right hands and distinguishes situations where there is no
press. For the purpose of training the neural network, a new dataset was created with diverse
hand images, as existing ones did not meet the needs of this work. The algorithm, developed
in the Python programming language, processes videos in real-time, collects fifteen
consecutive frames, and uses a neural network to recognize key presses. A graphical interface
allows interactive typing of letters, symbols, numbers and basic functions such as enter, space
and backspace. The model was trained and evaluated on a separate test set, with the best
model selected based on performance. Testing with three user groups demonstrated the
functionality and reliability of typing, with a drawback being the slow typing speed due to the
time requirements of the model and the camera