Poboljšani mikrogenetički algoritmi u upravljanju robotskim sustavima

doctoral thesis

• Author: Darko Dužanec
• Mentor: Zdenko Kovačić (mentor)
• Committee member: Stjepan Bogdan (predsjednik povjerenstva)
• Committee member: Darko Grundler (član povjerenstva)
• Committee member: Mario-Osvin Pavčević (član povjerenstva)
• Committee member: Marin Golub (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Control and Computer Engineering); Zagreb
• Defense date and country: 2010-10-14, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing; Process Computing
• Universal decimal classification (UDC): 004 - Computer science and technology. Computing. Data processing

Abstract

U ovom radu istražena je mogućnost primjene genetičkog algoritma za rješavanje optimizacijskih problema u realnom vremenu. Proveden je veliki broj numeričkih eksperimenata temeljem kojih je utvrđen utjecaj pojedinih parametara na ponašanje GA s ciljem utvrđivanja optimalnih parametara algoritma. U ovom radu predstavljena je metoda određivanja optimalnih parametara mikrogenetičkog algoritma temeljem poznatih optimalnih parametara genetičkog algoritma. Provedena je analiza ponašanja mikrogenetičkog algoritma pomoću Markovljevih lanaca. Temeljem te analize razvijen je operator kontrakcije prostora rješenja koji povećava točnost rješenja genetičkog algoritma. Opisani algoritam primijenjen je za rješavanje problema inverzne kinematike hiperredundantnog robota. Problem inverzne kinematike definiran je kao optimizacijski problem i definiran je vektorski oblik kromosoma pa su i genetički operatori prilagođeni vektorskom obliku kromosoma. Vektorski oblik genetičkog algoritma primijenjen je i za vremensku optimizaciju CAN sabirnice kojom se smanjuje vrijeme kašnjenja poruka kod većih opterećenja sabirnice. Mikrogenetički algoritam za rješavanje problema inverzne kinematike implementiran je na nekoliko računalnih platformi, dok je genetički algoritam za analizu CAN sabirnice testiran na postojećem upravljačkom sustavu vatrogasnog vozila.

Abstract (english)

This paper analyses the possibility of using genetic algorithms to solve real time optimization problems. A large number of experiments are performed to determine the influence of the parameters of the GA with the goal to determine the optimal set of parameters. A method is introduced to determine optimal parameters of microgenetic algorithms based on known parameters of the optimal genetic algorithm. An analysis of the microgenetic algorithm based on Markov chains is shown. The conclusions of this analysis are used to introduce the new solution space contraction parameter which increases the precision of the solution. The described algorithm is used to solve the inverse kinematics problem of a hiperredundant robot. The inverse kinematics problem is defined in form of an optimization problem and the chromosome is defined as a vector. The genetic operators are modified to operate with vector chromosomes. The vector form of the genetic algorithm is used to optimize the CAN bus protocol with the goal to reduce jitter at higher bus loads. The microgenetic algorithm for solving the inverse kinematics problem is implemented on several platforms and the genetic algorithm for CAN bus optimization is tested on an existing decentralized control system of an firefighting vehicle.

• Keywords: Genetički algoritam; mikrogenetički algoritam; optimiranje parametara; Markovljevi lanci; kontrakcija prostora rješenja; inverzna kinematika; hiperredundantni roboti; CAN sabirnica; optimizacija CAN sabirnice
• Keywords (english): Genetic algorithm; microgenetic algorithm; parameter optimization; Markov chains; solution space contraction; inverse kinematics; hyperredundant robots; CAN bus; CAN bus optimization
• Language: croatian
• URN:NBN: urn:nbn:hr:168:372168
• Project: Number: 036-0363078-3016; Title: Planiranje i izvođenje zadataka u robotskim i autonomnim sustavima; Leader: Stjepan Bogdan; Jurisdiction: Croatia; Funder: MZOS; Funding stream: ZP
• Project: Number: 036-0363078-3017; Title: Integrirano upravljanje robotskim sustavima u složenim okruženjima; Leader: Zdenko Kovačić; Jurisdiction: Croatia; Funder: MZOS; Funding stream: ZP
• Study programme: Title: Computer Science; Study programme type: university; Study level: postgraduate; Academic / professional title: Doktor znanosti (Doktor znanosti)
• Catalog URL: http://lib.fer.hr/cgi-bin/koha/opac-detail.pl?biblionumber=34151
• Type of resource: Text
• Extent: 177 str. ; 30 cm
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-03-19 09:25:10