機械手臂系統之適應性分散式控制器設計

Adaptive decentralized control of robot manipulators

許書豪(Su-Hau Hsu)

指導教授 ： 傅立成

國立臺灣大學/電機資訊學院/電機工程學研究所/博士(2004年)

https://doi.org/10.6342/NTU.2004.10246

摘要

在本論文中，我們利用高階輸入以及可變結構的設計觀念，針對機械手臂系統， 提出一個新型「適應性分散式控制器」的設計方法，以解決傳統上分散式控制器所遭遇的問題。由於該控制器是在分散式架構下運作，因此除了可減少通訊線路的鋪設 更可利用低階硬體完成實現，進而降低實作成本。尤其該控制器已搭配適應律的設計，因而具有處理參數未明之系統的能力。在實際應用時，毋需精準地知道系統參數，亦可達成控制目標，因此將會獲得許多的便利。
我們首先將這個適應性分散式控制器應用於解決「時變機械手臂」 (即機械手臂具由隨時間改變而變化的動力學參數)的控制問題。我們證明，無論機械手臂的參數如何快速變化，整個閉迴路系統的所有訊號將為有界，而且位置與速度追蹤誤差將指數收斂於一個小殘值區。值得注意的是，上述設計應用到「非時變機械手臂」 (即機械手臂具有固定大小的動力學參數)的控制問題，亦能獲得相同的控制性能。其次我們將這個適應性分散式控制器應用於解決「順從機械手臂」的控制問題，其中藉由環境撓性係數線上鑑別，進而完成目標軌跡的即時修正。我們證明不僅整個閉迴路系統的所有訊號將為有界之外，同時證明末端效應器的位置、姿態與接觸力追蹤誤差將漸進收斂於一個小殘值區。最後，我們將這個適應性分散式控制器應用於解決「電流饋送型感應馬達驅動之機械手臂」的控制問題。由於採用轉子磁通觀察器與針對感應馬達的設計適應律，我們證明，即使整個機電系統的參數出現不明的情況，不僅整個閉迴路系統的所有訊號均為有界，位置、速度與磁通追蹤誤差將漸進收斂於一個小殘值區。

關鍵字

機械手臂 ； 分散式控制 ； 適應性控制

並列摘要

In this thesis, a new approach of fully adaptive decentralized control is proposed for different tasks of robot manipulators. Since the proposed approach operates in a de-centralized manner, it is structurally simple such that only low-level hardware is re-quired in actual implementation. Because the adaptive mechanism is employed in the decentralized control scheme, the control performance could be guaranteed even if there is any parametric uncertainty in the robot manipulators systems. As a result, the proposed approach is very convenient for practical application.
This new adaptive decentralized controller is first applied to the time-varying robot manipulators, which is an important problem recently addressed in the filed of robot control. It is shown that, without any prior knowledge of the time-varying robot manipulators, the global stability of the closed-loop control systems could be achieved. Then, this new adaptive decentralized controller is applied to compliant robot ma-nipulators. After the on-line trajectory modification mechanism is provided, the force and position/orientation tracking performances of the end-effector are guaranteed even when the environment stiffness is unknown. Finally, this new adaptive decen-tralized controller is applied to the robot manipulators driven by current-fed induction motors. The servo and rotor-flux tracking performances is guaranteed here without any prior knowledge of the overall electromechanical systems.

並列關鍵字

Robot manipualtor ； decentralized control ； adaptive control

參考文獻

[1] B. Armstrong, O. Khatib, and J. Burdick, “The explicit dynamic model and iner-tial parameters of the PUMA 560 arm,” Proc. IEEE Int. Conf. Robot. Autom., San Francisco, USA, pp. 510-518, 1986.

[2] M. Bodson, J. Chiasson, and R. Novotnak, “High-performance induction motor control via input-output linearization,” IEEE Control System Magazine, Vol. 14, Issue 4, pp. 25-33, 1994.

[3] C. Canudas de Wit and B. Brogilato, “Direct adaptive impedance control in-cluding transition phases,” Automatica, Vol. 23, Issue 4, pp. 643-649, 1997.

[4] C. Canudas de Wit, R. Ortega, and S. I. Seleme, “Robot motion control using induction motor drives,” in Proc. IEEE Int. Conf. Robot and Automation, At-lanta, GA, pp. 533-538, 1993.

[5] S. Chiaverini and L. Sciavicco, “The parallel approach to force/position control of robotic manipulators,” IEEE Trans. Robot. Autom., vol. 9, pp. 361-373, 1993.