具關節力矩感測之撓性關節系統雙向控制
No Thumbnail Available
Date
2024
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
針對撓性關節系統,本研究旨在透過讀取關節力矩感測器和軸編碼器,獲得關節力矩與角位置,實現關節力矩控制,並應用於雙向控制。考慮關節撓性,對關節手臂進行系統鑑別之後,設計一雙迴路系統,內迴路實現關節力矩控制,而外迴路實現雙向控制,發展適用於撓性關節系統之雙向控制架構。此外,進一步設計一估測器,利用關節力矩和馬達端的位置,估測負載端的位置及外力,以免除負載位置編碼器的與外部力量感測器的安裝。實驗室內部成員共同研發設計與組裝一雙撓性關節系統,以作為實驗平台。該平台使用Sensodrive GmbH的SENSO-Joint 3008關節,利用RTX64即時作業系統將一般的個人電腦轉變為即時控制核心,以電腦為主站透過EtherCAT連接其他從站,並使用微軟的Visual Studio Code(VS)發展C++控制程式。本研究設計的方法將與其它控制方法進行實驗比較,包括不考慮關節撓性的控制方法及考慮撓性系統但未量測關節力矩的控制方法,通過性能指標來評估其優劣。實驗結果顯示,考慮關節撓性的控制系統優於基於剛體模型的控制系統,而量測關節力矩的回授控制又優於僅回授馬達及負載位置的控制。此外,估測器能準確地估測負載位置及外力,並應用於雙向控制,其表現優於其它控制方法,進一步提升系統性能。
This study focuses on flexiblejoint systems, achieving joint-torque control and applying it to bilateral control with joint torque sensors and shaft encoders. Considering joint flexibility, an inner loop is designed for torque control, while an outer loop achieves bilateral control. An estimator is developed using joint torque and motor position to accurately estimate load position and external force, eliminating the need for additional sensors.The experimental platform utilizes Sensodrive GmbH’s SENSO-Joint 3008 joints. The RTX64 real-time operating system is used to transform a general personal computer (PC) into a real-time controller. The PC connects to other devices via EtherCAT, and Visual Studio Code is employed for developing C++ control programs.The study experimentally compares various control methods, showing that systems considering flexibility and joint-torque feedback are superior to others. Experimental results show that the estimator can accurately estimate load position and external force, further enhancing system performance.
This study focuses on flexiblejoint systems, achieving joint-torque control and applying it to bilateral control with joint torque sensors and shaft encoders. Considering joint flexibility, an inner loop is designed for torque control, while an outer loop achieves bilateral control. An estimator is developed using joint torque and motor position to accurately estimate load position and external force, eliminating the need for additional sensors.The experimental platform utilizes Sensodrive GmbH’s SENSO-Joint 3008 joints. The RTX64 real-time operating system is used to transform a general personal computer (PC) into a real-time controller. The PC connects to other devices via EtherCAT, and Visual Studio Code is employed for developing C++ control programs.The study experimentally compares various control methods, showing that systems considering flexibility and joint-torque feedback are superior to others. Experimental results show that the estimator can accurately estimate load position and external force, further enhancing system performance.
Description
Keywords
撓性關節, 關節力矩感測, 雙向控制, 運動控制, 外力估測, flexible joint, joint torque sensing, bilateral control, motion control, external force estimation