| 摘要: | 本論文的研究目的是設計一利用遞迴式以小波函數為基礎之Elman類神經網路估測器之智慧型互補式滑動模態控制系統，控制X-Y-Ө三軸超音波線性馬達移動平台，以達到具有強健性之精密定位控制。本論文首先說明線型超音達馬達之工作原理，接著介紹線性超音波馬達之數學模型，因為其模型相當複雜，且馬達參數具有非線性且時變特性，易受溫度、負載轉矩及加在定子、轉子的彈簧靜壓力影響，故無法精確獲得。為了使X-Y-Ө三軸線性超音波馬達運動控制系統能在參數變化、摩擦力、外來干擾與多軸系統中交叉耦合干擾的影響下具備強健之控制性能，本論文依序提出Elman類神經網路控制系統、遞迴式以小波函數為基礎之Elman類神經網路控制系統、滑動模態控制系統、互補式滑動模態控制系統和智慧型互補式滑動模態控制系統，配合線上學習法則與滑動模態控制中之迫近控制律，分別控制X-Y-Ө三軸線性超音波馬達運動控制系統，以達到精密定位控制之目的。接著利用非均勻有理B-spline曲線插值法(non-uniform rational B-spline, NURBS)，針對XY軸設計了圓形和蝴蝶形輪廓軌跡命令以及Ө軸設計了弦波和梯形波輪廓軌跡命令，以驗證所發展智慧型控制架構之有效性。最後由不同軌跡追隨之實作結果證明相較於Elman類神經網路控制系統、遞迴式以小波函數為基礎之Elman類神經網路控制系統、滑動模態控制系統與互補式滑動模態控制系統，本論文所提出之智慧型互補式滑動模態控制系統具有最佳的控制性能與強健性。 An intelligent complementary sliding-mode control (ICSMC) system using a recurrent wavelet-based Elman neural network (RWENN) estimator is proposed to control the mover position of a linear ultrasonic motors (LUSMs)-based X-Y-Ө motion control stage and to achieve high precise position control with robustness in this study. First, the structure and operating principles of the LUSM were introduced. Since the motor parameters are highly non-linear and time-varying due to increase in temperature and change in operating conditions, the exact mathematical model of the LUSM is very difficult to obtain. Moreover, the control accuracy of the LUSM is influenced easily by the existence of uncertainties, which usually comprises system parameter variations, external disturbances, cross-coupled interference and friction force. In order to develop high performance and robust position control systems for LUSMs-based X-Y-Ө motion control stage under the occurrence of the uncertainties, five control systems including Elman neural network (ENN) control, recurrent wavelet-based Elman neural network (RWENN) control, sliding-mode control (SMC), complementary sliding-mode control (CSMC), and intelligent complementary sliding-mode control (ICSMC) systems, are proposed. Furethermore, to demonstrate the different control performances of various control systems, the circle, butterfly contours and sinusoid, trapezoid trajectories are designed for X-Y axes and Ө-axis, recepectively, using NURBS curve interpolator. Finally, some experimental results of various contours and trajectories tracking show that the proposed ICSMC owns the best control performance and robustness compared with the ENN, RWENN, SMC and CSMC systems. |
