Kognitivni model zatvorene okoline mobilnoga robota temeljen na mjerenjima

doctoral thesis

• Title (english): Cognitive model of closed environment of mobile robot based on measurements
• Author: Tomislav Pavlic https://orcid.org/0000-0001-5133-0665
• Mentor: Mladen Crneković (mentor)
• Committee member: Bojan Jerbić (predsjednik povjerenstva)
• Committee member: Tomislav Stipančić (član povjerenstva)
• Committee member: Zdenko Kovačić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Mechanical Engineering and Naval Architecture; Zagreb
• Defense date and country: 2021-10-07, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Mechanical Engineering
• Universal decimal classification (UDC): 004 - Computer science and technology. Computing. Data processing

Abstract

Posljednjih je godina u području mobilne robotike težište na metodama kod kojih se fuzijom mjerenja s više sustava dolazi do modela okoline koji su vjerojatnosnog karaktera. U ovom radu razvijen je novi pristup rješavanju SLAM problema, koji je temeljen na kognitivnom modelu zatvorene okoline mobilnog robota. Kognitivni model okoline je manje precizan od egzaktnog matematičkog, ali je nezaobilazan u interakciji robota s čovjekom.
Predloženo istraživanje bavi se razvojem modela učenja i planiranja djelovanja robota pri istraživanju nepoznatog prostora, na način sličan kako to i čovjek radi. Proces učenja temelji se na virtualnome dinamičkom modelu mobilnog robota, istovjetnom stvarnome robotu. Kretanje robota uvjetovano je razvijenim neuronskim mrežama te genetskim algoritmima.
U ovom radu predložena je nova metoda stvarnog (Sim-To-Real; STR) prijenosa znanja koja uključuje stadij vježbanja, stadij simultanog zaključivanja i stadij primjene naučenog na stvarnome robotu. Korištenjem jezika osnovnih kognitivnih elemenata opisana je okolina robota i njegov odnos prema toj okolini te je na temelju tog opisa dobivena viša razina informacija o stanju okoline u kojoj se robot nalazi, fuzijom mjernih podataka dobivenih iz različitih uređaja.
Video prezentacija istraživanja i eksperimenata dostupna je na sljedećoj poveznici:
https://youtu.be/KUxVIAe8hSM

Abstract (english)

This thesis is organized in 8 sections, as follows:
Section 1: Introduction. The first, introductory section, presents the motivation of the research, the overview of the area and the research so far. The main goal of the study is set, followed by a hypothesis. The plan and methodology of research work are presented. The rest of paper is organized as follows.
Section 2: Problem Description. In section 2, the primary goal of the research is established. Research work plan and research methodology are presented. Furthermore, the definition of tasks and terms are established.
Section 3: Mathematical Model of the Mobile Robot. In section 3, a definition of robot motion with mathematical models of the mobile robot system is described.
Section 4: Algorithms for Building a Cognitive Model of the eMIR Mobile Robot Environment. The cognitive model of the closed environment (developed algorithms for a cognitive behaviour of the mobile robot) is thoroughly described in section 4.
Section 5: Design of a Mechatronic System for Creating a Cognitive Model of Robot Space. In section 5, the development process of a mechatronic system used in research is described.
Section 6: Results. In section 6, experimental validation of the developed control model and the obtained results are described in detail.
Section 7: Discussion. The results obtained on the developed real mobile robot system were compared with those obtained by simulations. The applicability of the developed algorithms on a real robot system has been confirmed. This outcome is described in section 7.
Section 8: Conclusion. In section 8, an evaluation of the obtained results is discussed, and a critical review is given.
Appendix 1, contains information on additional materials such as codes, images, CAD models, simulation models, etc., and a link to a video of a cross-section of the entire study.
Appendix 2, on the other hand, contains the most important algorithms developed for the successful implementation of this research.

• Keywords: kognitivno ponašanje; strojno učenje; neuronska mreža; umjetna inteligencija; mobilni robot; percepcija okoline; kognitivni model zatvorenog okruženja
• Keywords (english): cognitive behaviour; machine learning; neural network; artificial intelligence; mobile robot; environment perception; cognitive model of closed environment
• Language: croatian
• URN:NBN: urn:nbn:hr:235:320631
• Study programme: Title: Mechanical Engineering and Naval Architecture; Study programme type: university; Study level: postgraduate; Academic / professional title: doktor/doktorica znanosti, područje tehničkih znanosti (doktor/doktorica znanosti, područje tehničkih znanosti)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2021-10-22 08:22:55