Sinteza trajektorije autonomnog pokretnog objekta oponašanjem operatera

doctoral thesis

• Title (english): Trajectory design for the autonomous moving vehicle by operater Imitation
• Author: Ranka Kulić
• Mentor: Zoran Vukić (mentor)
• Committee member: Ljubomir Kuljača (predsjednik povjerenstva)
• Committee member: Zoran Vukić (član povjerenstva)
• Committee member: Branko Novaković (član povjerenstva)
• Committee member: Nikola Bogunović (član povjerenstva)
• Committee member: Ivan Petrović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Control and Computer Engineering); Zagreb
• Defense date and country: 2004, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering
• Universal decimal classification (UDC): 621.3; APPLIED SCIENCES. MEDICINE. TECHNOLOGY; Electrical engineering

Abstract

Problem sinteze trajektorije robota koji se kreće u okolini u kojoj se nalaze nepokretne i pokretne prepreke se razmatra. Uopće, u literaturi je taj problem poznat kao problem planiranja trajektorije i kao problem izbjegavanja prepreka. U disertaciji se primjenjuje postupak oponašanja operatera sa ciljem da se projektira kontroler koji će u skladu sa usvojenim kriterijem optimalnosti voditi robot od početne do krajnje pozicije. Kod postupka oponašanja operatera strojni sustav uči na osnovu primjera koji su generirani dok robotom upravlja operater. Temeljem primjera dobijenih dok je robotom upravljao operater izbjegavajući nepokretnu pravokutnu prepreku, u ovom radu je pokazano da je, korištenjem više algoritama učenja, moguće sintetizirati pravila upravljanja, za situaciju sa skoro neograničenim brojem, kako konveksnih tako i konkavnih, i kako nepokretnih tako i pokretnih prepreka u ravnini. Predloženi postupak se pokazao primjenjivim i u 3-D prostoru. To ga suštinski razlikuje od drugih postupaka. Isto tako suštinska razlika predloženog postupka u odnosu na druge jest i u tom što je trajektoriju robota moguće definirati u ¨of – line¨ procesu, kao i što je sintetizirana pravila upravljanja moguće koristiti u ¨¨on – line¨ procesu bez uporabe složenih simboličkih modela okoline, odnosno, prepreka.

Abstract (english)

Dynamic path generation problem of robot in environment with other unmoving and moving objects is considered. Generally, the problem is known in literature as find path or robot motion planning and obstacle avoiding. In this paper it is applied the behavioral cloning approach to design the robot controller. In behavioral cloning, the system learns from control traces of a human operator. Taking in to account the examples which were obtained in training phase, when the robot was controlled by operator to avoid rectangle unmoving obstacle, some acceptable guidance rules were synthesized using different machine learning algorithms. Their importance is extended in order to be applied in some situation with infinite number of different unmoving and moving obstacles. That fact is a good reason for conclusion that an appropriate domain model were found. On the other hand, the exposed method gives results whose application in 3-D space is relatively simple. That fact substantially differs this methodology from the others. Also, the substantially advantage of this approach lies in the fact that a complete path can be defined off-line, without using sophisticated symbolical models of obstacles.

• Keywords: planiranje trajektorije robota; izbjegavanje prepreka; strojno učenje; oponašanje operatera; kriterij uspješnosti; linearna regresija; RBF neuronska mreža
• Keywords (english): robot motion planning; behavioral cloning; cloning success; obstacle avoiding; machine learning; linear regression; RBF neural network
• Language: croatian
• URN:NBN: urn:nbn:hr:168:416484
• Study programme: Title: Doctoral study programme "Electrical Engineering and Computing"; Study programme type: university; Study level: postgraduate; Academic / professional title: doktor/doktorica znanosti, po-dručje tehničkih znanosti (doktor/doktorica znanosti, po-dručje tehničkih znanosti)
• Type of resource: Text
• Extent: V, 184 str. ; 30 cm
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2020-03-03 13:14:31