本研究在Linux系統的Ubuntu20.04版本下透過機器人作業系統(Robot Operating System, ROS)開發控制軟體，並透過ROS點對點網路與其分散式架構將所有資訊進行資料傳遞並整合工業用IPC、六軸工業型機械手臂、2D相機及鏡頭、六軸力量/力矩感測器以及自適應夾爪等實現軟、硬體協同的設計。 本論文之任務目標在於透過導入影像視覺定位技術以及力量感測技術，開發一種能夠基於力量及視覺進行齒輪組裝的系統，本研究導入順應性力量控制(Compliance Force Control)的系統，來實現機械手臂進行齒輪組裝任務，並透過影像視覺定位技術，改善順應性力量控制在組裝過程中因位置的不確定性而導致組裝失敗的問題，在組裝前先透過影像視覺定位組裝物件，並將位置控制整合至順應性力量控制，最後透過實驗來組裝三種大小及外型不同的機車齒輪零件，來應證該系統的可行性，並由實驗結果得證該系統確實能夠用於齒輪箱組裝任務，此成果顯示本論文成功實現了透過力量及視覺的六軸機械手臂齒輪組裝系統。;This research is conducted on the Ubuntu 20.04 version of the Linux system, utilizing the Robot Operating System (ROS) for the development of control software. Through ROS′s point to-point networking and its distributed architecture, all information is transmitted and integrated, including industrial IPC, a six-axis industrial robotic arm, 2D cameras and lenses, a six-axis force/torque sensor, and an adaptive gripper. This achieves a design that synergizes both software and hardware components. The primary objective of this thesis is to develop a system for gear assembly based on force and vision using image-based visual localization and force sensing technologies. The research incorporates Compliance Force Control to enable the robotic arm to perform gear assembly tasks. Additionally, it addresses the issue of assembly failures due to positional uncertainties during the assembly process by incorporating image-based visual localization. Prior to assembly, the system employs image-based visual localization to locate the assembly objects and integrates position control into Compliance Force Control. Finally, experiments are conducted to assemble three different motorcycle gear components of varying sizes and shapes, demonstrating the feasibility of the system. The experimental results confirm that the system can indeed be used for gearbox assembly tasks. This achievement demonstrates the successful realization of a six-axis robotic arm gear assembly system through the integration of force and vision.
