Bešić, Sara (2019) Identifikacija dinamike laboratorijske grijalice zraka primjenom umjetne neuronske mreže. = System identification of laboratory fan heater using artificial neural network. Undergraduate thesis , Sveučilište u Zagrebu, Fakultet strojarstva i brodogradnje, UNSPECIFIED. Mentor: Brezak, Danko.
|
Text
Bešić_2019_zavrsni_preddiplomski.pdf - Published Version Jezik dokumenta:Croatian Download (1MB) | Preview |
|
![]() |
Text
Bešić_Sara_autorska_izjava_završni rad_2019.pdf - Published Version Restricted to Repository staff only Jezik dokumenta:Croatian Download (487kB) |
Abstract (Croatian)
Cilj umjetne inteligencije je ostvarenje sustava koji rješava probleme samostalno i inteligentno poput čovjeka. Kao što ideja o avionima potječe od promatranja ptica, tako je razvoj umjetnih neuronskih mreža potakunt spoznajama o građi i načinu funkcioniranja ljudskog mozga. U radu je ukratko objašnjen razvoj neurona i ideja neuronskih mreža te je opisana njihova struktura. Iako su još daleko od potpunog oponašanja funkcija mozga, neuronske mreže uspješno se primjenjuju na mnogim područjima. Prikazana je usporedba dva tipa neuronskih mreža, statičke i dinamičke mreže učene metodom povratnog prostiranja pogreške. Mreže su testirane i uspoređene na problemu identifikacije toplinskog procesa, realiziranog primjenom laboratorijskog modela grijalice zraka.
Abstract
The aim of artificial intelligence is to develop a system that solves problems independently and intelligently like a human. As the idea of a plane comes from observation of birds, so is the development of artificial neural networks enhanced by the knowledge of the structure and the way of functioning of the human brain. The paper briefly explains the development of neurons and ideas of neural networks, and also their structure is described. Although they are far from complete imitating brain functions, neural networks are successfully applied to many systems. Comparison of two types of neural networks, static and dynamic neural network trained with error-back propagation algorithm, is presented. Both networks were tested and compared on process identification problem realized using laboratory air heater testbed.
Item Type: | Thesis (Undergraduate thesis) |
---|---|
Uncontrolled Keywords: | umjetna inteligencija; neuron; umjetna neuronska mreža; statička neuronska mreža; dinamička neuronska mreža; povratno rasprostiranje pogreške |
Keywords (Croatian): | artificial intelligence; neuron; artificial neuron network; static neural network; dynamic neural network; error-back propagation |
Subjects: | TECHNICAL SCIENCE > Mechanical Engineering |
Divisions: | 900 Department of Robotics and Production System Automation > 910 Chair of Engineering Automation |
Date Deposited: | 27 Feb 2019 13:44 |
Last Modified: | 07 Nov 2019 16:52 |
URI: | http://repozitorij.fsb.hr/id/eprint/9207 |
Actions (login required)
![]() |
View Item |