VODENJE NELINEARNEGA MEHANIZMA Z METODAMI MEHKEGA RAČUNANJA

• Main Author: Oštir, Matjaž
• Other Authors: Gotlih, Karl, Rojko, Andreja
• Format: Bachelor Thesis
• Language: Slovenian
• Published: M. Oštir 2010
• Subjects: položajno vodenje mehanizmov; umetne nevronske mreže; mehka logika; nelinearne regulacije; mehko računanje; mehanizem z eno prostostno stopnjo; metoda izračunanega navora; position control of mechanisms artificial neural networks; fuzzy logic; nonlinear control; soft computing; mechanism with one degree of freedom; computed torque method; info:eu-repo/classification/udc/681.511.4:004.8(043.2); sami
• Online Access: https://dk.um.si/IzpisGradiva.php?id=15901; https://dk.um.si/Dokument.php?id=17732&dn=; https://plus.si.cobiss.net/opac7/bib/14631958?lang=sl

V diplomskem delu smo predstavili uporabo dveh mehkih metod računanja za vodenje nelinearnega mehanizma z eno prostostno stopnjo. Opisano je vodenje s pomočjo umetne nevronske mreže in mehke logike. Hkrati je bilo za primerjavo rezultatov obeh mehkih metod izvedeno vodenje s PD-regulatorjem. Sistemi vodenja so bili narejeni in simulirani v programskem paketu MATLAB/Simulink, sami algoritmi pa napisani v jeziku C. Izvajanje in testiranje algoritmov je potekalo na realnem mehanizmu s pomočjo DSP-2-modula. In this thesis, we present usage of two soft computing methods to control a nonlinear mechanism with one degree of freedom. Used soft computing methods are artificial neural networks and fuzzy logic. Obtained results are compared with those given by the PD controller. Control systems are made in the MATLAB/Simulink environment with algorithms written in C programming language. Implementation and testing was realised on a real mechanism by using DSP-2 module.