Repository of FERIT Osijek
Not a member yet
    4053 research outputs found

    Signal envelope filter

    No full text
    Zadatak ovog završnog rada je izrada DIY-sklopa filtra envelope signala te prikazivanje signala u vremenskoj domeni preko osciloskopa u laboratoriju i simulaciji ovisno o iznosu otpora potenciometra kao i mijenjanja stanja sklopke. U sadržaju je objašnjena pozadinska teorija oko envelope signala. Sklop je potrebno montirati i zalemiti pojedinačne komponente na mjesta određena PCB-om. Nakon izrade sklopa izvršava se simulacija sheme i laboratorijska mjerenja fizičkog sklopa. Osciloskopom se utvrdila značajna razlika u obliku signala u simulaciji i mjerenjima u laboratoriju zbog korištenja zamjenskih komponenata u simulaciji kao i simuliranje signala sa konstantnom amplitudom. Mjerenjem se utvrdilo frekvencijsko područje u kojem sklop ispravno radi.The task of this final paper is to solder a DIY envelope signal filter circuit and to display the signal in the time domain using an oscilloscope in the laboratory, while also simulating it depending on the resistance value of potentiometers and the position of the switches. The content includes the background theory of envelope filters. It was necessary to assemble and solder the individual components onto the designated places on the PCB. After completing the assembly, the circuit simulation and laboratory measurements of the physical circuit were carried out. The oscilloscope revealed a significant difference in the signal shape between the simulation and the laboratory measurements due to the use of substitute components in the simulation as well as simulating the signal with constant amplitude. The measurements determined the frequency range in which the circuit operates correctly

    Application of machine learning on microcontroller systems

    No full text
    Ovaj rad predstavio je jedan od načina implementacije modela strojnog učenja na mikroupravljač razvojne pločice Arduino Sense 33 BLE. Model koji je implementiran napravljen je da prepoznaje ključne riječi „crvena“ i „plava“, te da prepoznaje ostale zvukove i govor koje su kategoriziranje pod „tišina“. Kada su ključne riječi prepoznate na pločici se pali crveno ili plavo LED svjetlo. Kroz ovaj rad prikazuje se proces stvaranja modela za audio raspoznavanje koji je kreiran u platformi Edge impulse. Platforma je jednostavna i kompaktna za korištenje kao što je prikazano u radu. Kroz rad predstavljena je Edge impulse platforma te su obuhvaćene njene funkcionalnosti potrebne za stvaranje modela za audio prepoznavanje. Analiziran je model i njegova efikasnost i predstavljena je razlika između modela prije stavljanja na arduino mikroupravljač i poslije kako bi se prikazala razlika modela prije nego što je konvertiran u oblik koji se može koristiti ključne riječi (hrvatski ): Arduino IDE, Edge Impulse,LED, mikroupravljač, strojno učenjeThis paper presented one of the methods of implementing a machine learning model on the Arduino Sense 33 BLE development board. The implemented model was designed to recognize the keywords "red" and "blue", as well as to categorize other sounds and speech under "silence". When these keywords are recognized, a red or blue LED light is illuminated on the board. This paper demonstrates the process of creating an audio keyword recognition model using the Edge Impulse platform, which is shown to be user-friendly. The model and its effectiveness were analyzed, and the difference between the model before deployment on the Arduino microcontroller and after is presented to illustrate the transformation of the model into a deployable format

    Designing an energetically self-sufficient residential building

    No full text
    Projektiranjem energetski samodostatnog stambenog objekta analiziraju se izazovi i mogućnosti primjene solarne energije i drugih tehnologija za postizanje energetske neovisnosti. S obzirom na sve veće energetske potrebe i utjecaj klimatskih promjena, energetska samodostatnost postaje ključni cilj. Najistaknutija je potreba za učinkovitijim korištenjem obnovljivih izvora energije, s posebnim naglaskom na solarnu energiju i dizalice topline. Cilj ovog rada je dizajnirati sustav za energetski neovisni stambeni objekt koji koristi solarne panele za proizvodnju električne energije i dizalicu topline za grijanje i hlađenje objekta. Stambeni objekt je projektiran prema važećim zakonima i standardima u Hrvatskoj, s naglaskom na energetski učinkovit dizajn i održivost. U radu je analizirana potrošnja energije te je projektiran fotonaponski sustav za pokrivanje energetskih potreba. Korišteni su alati kao što je PVGIS za procjenu proizvodnje energije iz solarnih panela. Analizirani su i stvarni podaci o potrošnji i proizvodnji električne energije za postojeći objekt, što omogućava optimizaciju sustava. Na temelju rezultata zaključeno je da je moguće postići energetsku neovisnost pravilnim dimenzioniranjem i projektiranjem sustava za obnovljive izvore energije, a budući aspekti ovog projekta mogli bi se usmjeriti na daljnja istraživanja i unapređenje postojećih metoda za povećanje učinkovitosti fotonaponskih sustava i dizalica topline.The design of an energy self-sufficient residential building analyzes the challenges and possibilities of using solar energy and other technologies to achieve energy independence. Given the growing energy demands and the impact of climate change, energy self-sufficiency has become a key goal. The most prominent need is for more efficient use of renewable energy sources, with a focus on solar energy and heat pumps. The aim of this paper is to design a system for an energy-independent residential building that utilizes solar panels to generate electricity and a heat pump for heating and cooling. The residential building is designed according to the applicable laws and standards in Croatia, with an emphasis on energy-efficient design and sustainability. The paper analyzes energy consumption and designs a photovoltaic system to meet the energy needs. Tools such as PVGIS were used to estimate the energy production from the solar panels. Actual data on energy consumption and production for the existing building were also analyzed, allowing for system optimization. Based on the results, it was concluded that energy independence can be achieved with proper sizing and design of renewable energy systems. Future aspects of this project could focus on further research and improvement of existing methods to enhance the efficiency of photovoltaic systems and heat pumps. keywords: energy self-sufficient residential building, photovoltaic system, solar power plant, electrical energy, heat pum

    A web application for renal disease risk assessment and monitoring based on machine learning

    No full text
    U završnom radu analizirani su problemi i izazovi u procjeni rizika od bubrežnih bolesti te praćenju njihova tijeka, kao i stanje u tom području i postojeća slična rješenja. Također je bilo važno utvrditi značajke modela i odgovarajuće postupke strojnog učenja prema vrsti problema, pripremiti odgovarajuće skupove podataka za treniranje i testiranje modela, obaviti treniranje i evaluaciju modela strojnog učenja te ga pripremiti za integraciju u web sustav s mogućnošću generiranja preporuka temeljenih na stanju pacijenta. Nadalje, definirani su funkcionalni i nefunkcionalni zahtjevi te model i arhitektura web aplikacije s aspekta klijentske i poslužiteljske strane. Web aplikacija je trebala omogućiti registraciju i prijavu korisnika, izradu osobnog profila, unos i pohranu podataka, procjenu rizika, praćenje bolesti te generiranje i prikaz preporuka pacijentima. Razvoj web aplikacije i modela strojnog učenja proveden je korištenjem prikladnih programskih jezika, tehnologija i razvojnih okvira, a testirani su i analizirani na relevantnim ulaznim podacima i slučajevima korištenja.The final thesis analyzed the problems and challenges in assessing the risk of kidney diseases and monitoring their progression, as well as the state of this field and existing similar solutions. It was also important to determine the characteristics of the model and appropriate machine learning procedures according to the type of problem, prepare suitable datasets for training and testing the model, conduct training and evaluation of the machine learning model, and prepare it for integration into a web system with the capability of generating recommendations based on the patient’s condition. Furthermore, functional and non-functional requirements were defined, along with the model and architecture of the web application from the perspective of the client and server side. The web application was intended to allow user registration and login, creation of a personal profile, data entry and storage, risk assessment, disease monitoring, and the generation and display of recommendations for patients. The development of the web application and machine learning model was carried out using appropriate programming languages, technologies, and development frameworks, and they were tested and analyzed on relevant input data and use cases

    The role of the aspect-oriented paradigm (AOP) in application development

    No full text
    Cilj ovoga rada je opisati pojam aspektno orijentirane paradigme, specifične probleme koje koncept adresira i primjerom pokazati ulogu iste u razvoju aplikacije. U radu je opisano što je AOP i koje su njegove prednosti odnosno njegova poboljšanja i načela. Obrađeni su dominantni okviri AOP-a na način da su opisani AspectJ, Spring i PostSharp razvojni okviri i uspoređeni na temelju 10 kategorija kao što su implementacija, integracija, performanse itd. Opisan je pojam aspekta kao ključne jedinice modularnosti koja omogućuje dodavanje funkcionalnosti bez kompliciranja osnovnog koda. Opisane su OOP i AOP paradigme, njihove uloge, prednosti, nedostatci te njihova povezanost. U završnom dijelu rada na praktičnom primjeru prikazano je kako uz PostSharp razvojni okvir adresirati razne aspekte AOP paradigme koristeći aspektne klase i atribute u C# Web aplikaciji.The goal of this paper is to describe the concept of an aspect-oriented paradigm, the specific problems that the concept addresses and to demonstrate its role in the development of applications with an example. The paper describes what AOP is and what its advantages are, i.e. its improvements and principles. The dominant frameworks of AOP are covered in such a way that AspectJ, Spring and PostSharp development frameworks are described and compared based on 10 categories such as implementation, integration, performance, etc. The concept of an aspect is described as a key unit of modularity that allows adding functionality without complicating the basic code. The OOP and AOP paradigms, their roles, advantages, disadvantages and their connection are described. In the final part of the work, a practical example shows how to address various aspects of the AOP paradigm using the aspect class and attributes in a C# Web application with the PostSharp development framework

    Analysis of increasing the integration of different types of RES on microgrid adequacy

    No full text
    U radu je opisana analiza adekvatnosti mikromreže provedena u Digsilent PowerFactory programskom paketu koja pruža ključne uvide u ponašanje sustava s fotonaponskim izvorima i vjetrogeneratorima. U radu su opisane mikromreže te utjecaj obnovljivih izvora energije na mikromrežu i provedena je analiza adekvatnosti mikromreže za različite obnovljive izvore energije. Promjene vjerojatnosti stanja rada i stanja kvara pri analizi ukazuje na izazove u održavanju pouzdanosti mikromreže. Rezultati istraživanja, prikazani slikama i dijagramima, pružaju važan uvid u dinamiku mikromreže, a preporučuje se daljnje istraživanje s fokusom na strategije za smanjenje LOLP-a radi održavanja stabilnosti sustava.The paper describes the adequacy analysis of a microgrid conducted in the DigSilent PowerFactory software package, providing key insights into the behavior of the system with photovoltaic sources and wind generators. The paper outlines microgrids and the impact of renewable energy sources on the microgrid, conducting an adequacy analysis for various renewable energy sources. Changes in the probability of operating states and fault conditions during the analysis indicate challenges in maintaining the reliability of the microgrid. The research results, presented through images and diagrams, offer significant insights into the dynamics of the microgrid, with a recommendation for further research focusing on strategies to reduce Loss of Load Probability (LOLP) for sustaining system stability

    Application of augmented reality in iOS game development

    No full text
    U diplomskom radu istražena je proširena stvarnost, mobilne računalne igre, povezanost proširene stvarnosti i mobilnih računalnih igara te trenutno stanje korištenja proširene stvarnosti u mobilnim računalnim igrama. Također, opisani su izazovi i pristupi pri razvoju mobilne računalne igre u proširenoj stvarnosti. Programski ostvarena mobilna iOS računalna igra omogućuje primjenu proširene stvarnosti, stoga korisnici imaju mogućnost igrati poznatu arkadnu igru Snake u proširenoj stvarnosti. Rješenje mobilne računalne igre pisano je prema predlošku arhitekture VIPER, a metodologija izrade oslanjala se na reaktivno i asinkrono programiranje. Također, opisan je utjecaj arhitekture VIPER pri razvoju mobilne računalne igre. Mobilna računalna igra Snake omogućuje korisnicima upravljanje objektima u proširenoj stvarnosti prema pravilima igre, rangiranje igrača prema ostvarenim rezultatima i pohranu rezultata. Ispitivanje rada aplikacije pokazalo je njeno očekivano ponašanje, a iz istraživanja korisničkog iskustva saznaje se da je utjecaj proširene stvarnosti na korisnike pozitivan.The aim of this master's thesis was to explore augmenter reality, mobile games, connection between augmented reality and mobile games, and the current state of using augmented reality in mobile games. Furthermore, this thesis describes challenges and approaches in developing mobile games in augmented reality. The software-realized mobile iOS game enables the application of augmented reality, so users have the ability to play the famous arcade game Snake in augmented reality. The mobile game solution was written according to the VIPER architecture template, and the development methodology was based on reactive and asynchronous programming. Additionally, the thesis describes the influence of the VIPER architecture in the development of a mobile game. The mobile game Snake allows users to manage objects in augmented reality according to the rules of the game, rank players according to the achieved results and store the results. Testing the operation of the application showed its expected behavior, and from the user experience research it was found that the impact of augmented reality on users is positive

    Anomaly Detection in Network Traffic Using Machine Learning Classification in a Hybrid Software-Defined Network

    No full text
    Disertacija predstavlja inovativan pristup kibernetičkoj sigurnosti u kontekstu hibridnih programski definiranih mreža (SDN), s naglaskom na integraciju tehnika strojnog učenja za otkrivanje anomalija u mrežnom prometu. Istraživanje je usmjereno na rješavanje sigurnosnih izazova u SDN-arhitekturama kroz razvoj sustava za otkrivanje napada (NIDS) temeljenog na strojnom učenju, koji omogućava učinkovito otkrivanje i odgovor na prijetnje u stvarnom vremenu. Predloženi model kombinira prednosti tradicionalnih mrežnih arhitektura s prilagodljivošću SDN-a, pružajući skalabilno i učinkovito otkrivanje anomalija. Postupak odabira i prilagodbe ulaznih podataka, zajedno s optimizacijom algoritma klasifikacije, omogućava visoku točnost modela, što je vrlo važno za zaštitu kritičnih segmenata IT-sustava u korporativnoj mreži. Disertacija također uključuje detaljan pregled postojeće literature, istraživanje metoda integracije strojnog učenja sa SDN-om te rezultate verifikacije modela u stvarnom mrežnom okruženju. Dodatni istraživački doprinos uključuje razvoj metode verifikacije temeljene na analizi višestrukih NetFlow-zapisa, čime se dodatno potvrđuje učinkovitost predloženog modela. Znanstveni doprinosi ove disertacije uključuju: 1. Postupak odabira i prilagodbe skupa ulaznih podataka za algoritam klasifikacije radi otkrivanja anomalija u hibridnoj programski definiranoj mreži. 2. Predložen je model za otkrivanje anomalija u mrežnom prometu hibridne SDN-mreže, koji koristi algoritme strojnog učenja. 3. Osmišljena je metoda za verifikaciju predloženog modela u stvarnom mrežnom okruženju. Disertacija donosi značajan doprinos području mrežne sigurnosti, s posebnim naglaskom na napredna i inteligentna sigurnosna rješenja unutar hibridnih SDN-arhitektura.This dissertation presents an innovative approach to cybersecurity in the context of hybrid Software-Defined Networks (SDN), with a focus on integrating machine learning techniques for anomaly detection in network traffic. The research is aimed at addressing security challenges in SDN architectures through the development of a Network Intrusion Detection System (NIDS) based on machine learning, enabling effective real-time threat recognition and response. The proposed model combines the strengths of traditional network architectures with the flexibility and adaptability of SDN, providing scalable and efficient anomaly detection. The process of selecting and adapting input data, along with the optimization of the classification algorithm, enables high model accuracy, which is crucial for protecting critical segments of IT systems within corporate networks. The dissertation also includes a detailed review of existing literature, an exploration of methods for integrating machine learning with SDN, and results from the model’s verification in a real network environment. An additional research contribution includes the development of a verification method based on the analysis of multiple NetFlow records, further confirming the effectiveness of the proposed model. The scientific contributions of this dissertation include: 1. The development of a process for selecting and adapting the input data set for the classification algorithm to detect anomalies in a hybrid Software-Defined Network. 2. A proposed model for anomaly detection in the network traffic of a hybrid SDN, utilizing machine learning algorithms. 3. The design of a method for verifying the proposed model in a real network environment. The dissertation makes a significant contribution to the field of network security, with a particular emphasis on advanced and intelligent security solutions within hybrid SDN architectures

    Computer simulation of broadband network access.

    No full text
    U radu je dan pregled različitih tehnika multipleksiranja u bežičnim komunikacijskim sustavima, s naglaskom na TDMA, FDMA i CDMA tehnologije. Analizirani su procesi modulacije, širenja i demodulacije signala, te je prikazan njihov utjecaj na kvalitetu i efikasnost komunikacije. Simulirani su različiti scenariji koji demonstriraju kako ove tehnike upravljaju šumom i interferencijom u realnim komunikacijskim kanalima. Rezultati simulacija daju uvid u njihovu primjenu u modernim telekomunikacijskim mrežama.This paper provides an overview of various multiplexing techniques in wireless communication systems, focusing on TDMA, FDMA, and CDMA technologies. The processes of modulation, spreading, and demodulation of signals are analyzed, showcasing their impact on the quality and efficiency of communication. Various scenarios have been simulated to demonstrate how these techniques manage noise and interference in real communication channels. The results of the simulations offer insights into their application in modern telecommunication networks

    Web application for keeping movie notes and reviewing movies

    No full text
    U današnjem svijetu gdje se broj sadržaja koje je moguće konzumirati svakodnevno povećava, javlja se potreba za alatima koji mogu u tome pomoći. Iako mnoga rješenja za problematiku vođenja evidencije i recenzije filmova već postoje, varijacije u detaljima među njima mogu činiti veliku razliku nekim korisnicima. Stoga uvijek postoji potreba za dodatnim rješenjima na tržištu. Cilj ovoga završnoga rada bio je prikazati postupak izrade programskog rješenja za traženi problem, uz objašnjenje tehnologija koje su korištene, kao i razloge zbog kojih su odabrane. U razvoju aplikacije korištene su tehnologije Django, Bootstrap, HTML te CSS. Aplikacija omogućava korisniku izradu korisničkog računa te prijavu i odjavu. Korisnik može pretraživati filmove i vidjeti podatke o njima. Filmove koje je već pogledao može dodati u listu pregledanih filmova. Filmove koje želi pogledati može dodati u zasebnu listu. Korisniku je dan prikaz drugih korisnika aplikacije i dan mu je pregled filmova koje su pogledali. Svaki korisnik može promijeniti svoje korisničke podatke.In today's world, where the amount of content that can be consumed increases daily, there is a need for tools that can aid in solving this problem. Although many solutions to the problem of record keeping and reviewing movies already do exist, the variations in detail between them can make a big difference to some users. Therefore, there is always a need for additional solutions on the market. The goal of this final thesis was to present the process of creating a software solution for the requested problem, along with an explanation of the technologies that were used, as well as the reasons why they were chosen. The application was developed using technologies such as Django, Bootstrap, HTML and CSS. The application allows the user to create a user account, as well as log in and log out. The user can search for movies and see information about them. The user can add movies he has already watched to the list of watched movies. He can add the movies he wants to watch to a separate list. The user is given a view of other users of the application and is given an overview of the movies they have watched. Every user can change their user data. keywords: application, Django, tracking, movie

    103

    full texts

    4,053

    metadata records
    Updated in last 30 days.
    Repository of FERIT Osijek
    Access Repository Dashboard
    Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇