Načrtovanje samonastavljivega regulatorja 2 DOF robota s pomočjo BA algoritma

diplomsko delo

Dušan Fister (Author), Miran Brezočnik (Mentor), Riko Šafarič (Mentor), Iztok Fister (Co-mentor)

Abstract

Algoritmi po vzoru iz narave dandanes pokrivajo več raziskovalnih področij. Aplikacije, s katerimi rešujemo različne optimizacijske probleme v industriji in drugih področjih človekove dejavnosti, povečujejo kakovost proizvoda, zmanjšujejo časovni okvir načrtovanja ali kako drugače lajšajo reševanja problema. Naš problem predstavlja načrtovanje parametrov položajnega regulatorja na dvoosnem robotskem mehanizmu s pomočjo algoritmov po vzoru iz narave. Pri tem med seboj primerjamo genetski algoritem ter algoritem po vzoru obnašanja netopirjev. Spoznamo in potrdimo osnovne značilnosti obeh algoritmov, ju preizkusimo na realni laboratorijski aplikaciji, ter ugotovimo, da algoritem po vzoru obnašanja netopirjev izboljšuje rezultate genetskega algoritma na našem problemu. S tem smo prišli do sklepa, da lahko algoritme po vzoru iz narave uspešno uporabimo za reševanje realnih problemov iz prakse.

Keywords

algoritmi po vzoru obnašanja netopirjev;mehatronika;regulacije;

Data

Language:	Slovenian
Year of publishing:	2015
Typology:	2.11 - Undergraduate Thesis
Organization:	UM FS - Faculty of Mechanical Engineering
Publisher:	[D. Fister]
UDC:	004.021:681.51(043.2)
COBISS:	19156246
Views:	1442
Downloads:	97
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Self-tuning controller design of 2 DOF robot using BA algorithm
Secondary abstract:	Nature-inspired algorithms present a bright star of researching. Implemented on different optimizational, industrial or researchable applications they offer an excellent opportunity to increase quality of product, reduce time needed for design or ease the procedure of solving modern problems. Our main topic presents the 2 dof robotic mechanism, equipped with nature-inspired algorithms in order to design self-tuning controller. We tested and compared two algorithms - Genetic algorithm and Bat algorithm, met their basics of working and according to real laboratory tests confirmed the execution of optimizational process. Bat algorithm did prove better working in the laboratory experiments. We conclude, that it was worth to use the nature-inspired algorithms for our problem.
Secondary keywords:	nature-inspired algorithms;mechatronics;controlling;
URN:	URN:SI:UM:
Type (COBISS):	Bachelor thesis/paper
Thesis comment:	Univ. v Mariboru, Fak. za strojništvo
Pages:	XV, 43 str.
ID:	8888258