Robotska montaža u nesređenoj radnoj okolini

Grolinger, Katarina (1997) Robotska montaža u nesređenoj radnoj okolini. = Robotic assembly in unpredictable working environment. Scientific master's thesis , Sveučilište u Zagrebu, Fakultet strojarstva i brodogradnje, UNSPECIFIED. Mentor: Jerbić, Bojan.

Full text not available from this repository.

Abstract (Croatian)

Montaža je područje koje ima znatan udio u troškovima i vremenu proizvodnje. Stoga su današnji napori usmjereni k unapređenju montaže, posebno fleksibilne montaže uz primjenu robota. Ovaj rad se bavi razvojem inteligentnog robotskog sustava za primjenu u montaži, koji je sposoban rješavati različite zadaće prilagođavajući svoje ponašanje promjenljivom okolišu, uključujući snalaženje u neizvjesnim i nepredvidivim situacijama. Da bi to postigao robot mora moći naći rješenje u nepoznatoj situaciji, učiti ponašanje, što znaci procedure ponašanja zajedno s odgovarajućim znanjem o strukturi radnog prostora, i prepoznati radni prostor. Problem inteligentnog robotskog ponašanja se suočava s nepredvidivo velikim brojem varijanti problema što ne može biti obuhvaćeno egzaktnom domenom znanja, stoga se razvoj zasniva na aktivnim metodama učenja. Sustav primjenjuje metodu prisilnog učenja baziranu na strategijskim i slučajnim pokušajima za pronalaženje rješenja problema i neuronsku mrežu za pamćenje i prepoznavanje struktura radnog prostora. Razvijena je adaptivna neizrazita osjenčana (A\NO) neuronska mrena. Njena struktura je inspirirana teorijom adaptivne rezonancije, premda ima dodatnu osjenčanu skrivenu razinu, specifično pravilo učenja i fazu inicijalizacije.

Abstract

The assembly is the field which takes great part of time and costs of production. So today efforts are turned toward improvement off assembly, particularly flexible assembly with robot utilization. This paper is concerned with the design of intelligent robotic system for use in assembly conditions, which is able to solve various tasks while adopting its behavior to the variable environment, including handling uncertain and unexpected situations. To achieve this the robot has to be able to find the solution in unknown situations, to learn experienced knowledge, that means action procedure together with corresponding knowledge on the work space structure, and to recognize working environment. The problem of intelligent robot behavior copes with unpredictable large number of problem varieties which can not be captured in exact knowledge domain, hence its development is based on the active learning methods. The system implements the reinforcement learning method based on strategic and random attempts for fmding problem's solutions and neural network for memorizing and recognizing work space structure. The Adaptive Fuzzy Shadowed (ANO) neural network is developed. It's architecture is inspired with Adaptive Resonance Theory, though it has the additional shadowed hidden layer, specific learning rule and initialization phase.

Item Type: Thesis (Scientific master's thesis)
Uncontrolled Keywords: robotska montana; učenje; inteligencija; neuronska mreža
Keywords (Croatian): robotic assembly; learning; intelligence; neural network
Divisions: 900 Department of Robotics and Production System Automation > 920 Chair of Manufacturing and Assembly System Planning
Date Deposited: 22 Sep 2014 18:00
Last Modified: 16 Oct 2015 13:20
URI: http://repozitorij.fsb.hr/id/eprint/113

Actions (login required)

View Item View Item

Nema podataka za dohvacanje citata