Masinnägemise ja ROS MoveIt-i integreerimine robot manipulaatori täpseks juhtimiseks
Laen...
Kuupäev
2025
Kättesaadavus
10.09.2025
Autorid
Ajakirja pealkiri
Ajakirja ISSN
Köite pealkiri
Kirjastaja
Eesti Maaülikool
Abstrakt
Käesolev töö on osa AIRE demoprojektist, mille eesmärk on arendada süsteem detailide rakiseta paigutamiseks kordinaatmõõtepingile. Ning integreerida UR3 koostöörobot, RealSense sügavuskaamerad ja ROS 2 MoveIt! 2 platvorm ühtseks süsteemiks detailide paigutuseks.
Andmed koguti YOLOv11 algoritmil põhineva masinnägemise ja sügavuskaamerate abil, süsteemi arhitektuur loodi ROS 2 platvormil
Õnnestus liidestada kõik komponendid ja määrata objektide koordinaadid, kuid roboti tegelik liikumine jäi tarkvaraprobleemide tõttu saavutamata. Töö näitas, et UR3 juhtimine ROS 2 ja MoveIt 2 kaudu vajab täiendavat arendust, kuid loodud lahendus on alus edasisteks arendusteks ja tööstuslikuks rakendamiseks.
This thesis is part of the AIRE demo project focused on developing a system for fixtureless part placement onto a measurement surface. The aim of the thesis is to integrate the UR3 collaborative robot, RealSense depth cameras, and the ROS 2 MoveIt platform into a unified system for precise object placement. Data was collected using YOLOv11-based machine vision and depth cameras; the system was built on the ROS 2 platform, including hand-eye calibration and motion planning. While all components were successfully integrated and object coordinates determined, the robot's actual movement via MoveIt! 2 failed due to software issues. The results show that UR3 control via ROS 2 and MoveIt 2 requires further development, but the solution provides a solid foundation for future work and industrial applications.
This thesis is part of the AIRE demo project focused on developing a system for fixtureless part placement onto a measurement surface. The aim of the thesis is to integrate the UR3 collaborative robot, RealSense depth cameras, and the ROS 2 MoveIt platform into a unified system for precise object placement. Data was collected using YOLOv11-based machine vision and depth cameras; the system was built on the ROS 2 platform, including hand-eye calibration and motion planning. While all components were successfully integrated and object coordinates determined, the robot's actual movement via MoveIt! 2 failed due to software issues. The results show that UR3 control via ROS 2 and MoveIt 2 requires further development, but the solution provides a solid foundation for future work and industrial applications.
Kirjeldus
Rakenduskõrgharidusõppe lõputöö
Tehnotroonika õppekaval
Märksõnad
lõputööd, masinnägemine, ROS 2, robootika, MoveIt! 2