Razvoj krmiljenja mobilnega robota z metodo vzpodbujevalnega učenja

Planinšek, Nejc (author), Vrabič, Rok (mentor)

Strojno učenje se vse pogosteje pojavlja kot rešitev problemov, ki jih je težko rešiti s klasičnimi pristopi. Pri strojnem učenju algoritem naučimo na osnovi podatkov, namesto da bi ga eksplicitno napisali. V tej nalogi smo z metodami strojnega učenja razvili krmilnik mobilnega robota za sledenje črti. Razvili smo zaznavalo črte, ki deluje odlično v različnih pogojih. Ustvarili smo simulacijsko okolje za preizkušanje krmiljenja robota in v njem z metodami vzpodbujevalnega učenja razvili krmilnik, ki sledi črti. Krmilnik smo iz simulacije prenesli v realni svet. V simulaciji krmilnik deluje primerljivo s klasičnim PID krmilnikom, v realnem svetu pa precej slabše, kar bi lahko izboljšali v nadaljnjem delu. Zaznavalo črte in krmilnik sta sposobna realnočasovnega delovanja na omejenih platformah, kot je Raspberry Pi.

• Language: Slovenian
• Keywords: strojno učenje, vzpodbujevalno učenje, umetne nevronske mreže, simulacijsko okolje, sledenje črti, mobilni roboti
• Work type: Master's thesis
• Typology: 2.09 - Master's Thesis
• Organization: FS - Faculty of Mechanical Engineering
• Place of publishing: Ljubljana
• Publisher: [N. Planinšek]
• Year: 2019
• Number of pages: XXI, 58 str.
• PID: 20.500.12556/RUL-107876
• UDC: 007.52:004.83(043.2)
• COBISS.SI-ID: 16656155
• Publication date in RUL: 01.06.2019

PLANINŠEK, Nejc, 2019, Razvoj krmiljenja mobilnega robota z metodo vzpodbujevalnega učenja [online]. Master’s thesis. Ljubljana : N. Planinšek. [Accessed 31 March 2025]. Retrieved from: https://repozitorij.uni-lj.si/IzpisGradiva.php?lang=eng&id=107876

Secondary language

• Language: English
• Title: Design of mobile robot control using reinforcement learning methods

Abstract

Machine learning is increasingly emerging as a solution to problems that are difficult to solve with classical approaches. In machine learning, the algorithm is trained on data, rather than written explicitly. In this thesis, we developed a mobile robot controller using methods of machine learning. We developed a line detector, that works well under different conditions. We created a simulation environment, designed for testing robot control algorithms, and developed a line following controller using reinforcement learning methods. The controller was transferred from simulation to the real world. In the simulation, controller performance is comparable to the classic PID controller, whereas in the real world, it is considerably worse. That could be improved in future work. The detector and controller are capable of real-time operation on limited platforms such as Raspberry Pi.

• Keywords: machine learning, reinforcement learning, artificial neural networks, simulation enviornment, line following, mobile robots