機器臂以交流感應馬達驅動之位置適應控制

Abstract

在本論文中我們針對既有的文獻中以交流感應馬達為致動器之單一連桿機械臂系統 所提出的非線性控制器做一番簡單的整理與敘述。並且對此非線性控制器的性能與 系統穩定性藉由模擬與實驗結果說明了此方法所設計之控制器性能優越。我們也對 如何建立整個系統的實驗架構有完整的說明。由於此非線性控制器是在系統模型參 數均正確得知的情況下所設計的，亦即並未考慮系統模型參數的不確定性所造成的 影響。所以我們亦針對機械子系統模型參數的不確定性加以考慮，並提出新的轉子 磁通與定子電流觀測器來發展適應性、部分狀態迴授之位置追蹤控制器來補償因未 考慮系統模型參數的不確定性使得位置追蹤性能降低的現象。 我們利用非線性backstepping系統化的設計方法配合nonlinear damping來保證系統的穩 定性。在機械子系統含有不確定性的模型參數之下，此電壓輸入之控制器只需量測 連桿位置與定子電流訊號，由模擬結果可知連桿追蹤誤差量為漸近穩定而且轉子磁 通與定子電流估測值亦為漸近穩定。透過單一連桿機械臂實驗系統之結果亦說明此 法之可行性。

In this thesis, we have made plainly delineation for the proposed nonlinear voltage input controller for robot manipulators driven by induction motors in the existent literature. And follow the same design procedure, simulated and experimental result is then presented to illustrate the position tracking performance for the proposed controller and system stability. Specially, we make complete representation for how to setup the experimental hardware. Owing to the nonlinear controller is designed under the assumption that exact model knowledge is known, say, without considering the effect of system model parametric uncertainties. Then, we focused on the mechanical subsystem parametric uncertainties and proposed novel rotor flux and stator current observer to derive the adaptive, partial-state feedback, position tracking controller to compensate the degraded performance caused by without considering the system model parametric uncertainties. Through the use of systematic design approach backstepping and nonlinear damping to guarantee boundedness of trajectories even when no upper bound on the uncertainties is known. Under the mechanical subsystem parametric uncertainties, the proposed voltage input controller only requires measurements of link position and stator winding currents. Finally, we conclude a theorem to insure the controller stability and simulation results are presented to illustrate the link position tracking error, and rotor flux and stator current estimated error are asymptotically stable.