基於教與學最佳化策略之適應性混合模糊PID控制應用於音圈馬達運動平台

Abstract

本論文目標為針對音圈馬達運動平台設計一適應性混合模糊比例-積分-微分控制策略，使該平台具備優異之定位精度與強健性能。首先說明音圈馬達運動平台的系統架構及運作原理，經由系統鑑別推導出馬達數學模型以及系統參數，將回授訊號達到或保持在理想值使系統變得更加準確且穩定。接著，以模糊理論設計一個模糊PID(Fuzzy Proportional–Integral–Derivative, FPID)控制器，透過動態調整控制增益的方式改善系統穩定度，進一步提升動態響應和強健性。之後，為了進一步提升系統的抗干擾能力，本研究設計一個基於教與學演算法最佳化模糊歸屬函數的適應性混和模糊控制器，讓控制器能夠隨著輸入誤差動態調整歸屬函數的區間，使模糊系統在相同誤差下能反應出更精確的歸屬度，解模糊化得到前饋控制力將進一步提高系統的穩定度並抑制外部干擾的影響。本論文以數位訊號處理器實現上述控制策略並比較兩種追蹤軌跡，最後由實驗結果得知最佳化模糊歸屬函數的適應性混和模糊控制器相比於傳統PID控制器的控制性能，加入雜訊的窗形軌跡平均誤差改善58.28 %，加入雜訊的花瓣形軌跡平均誤差改善66.32 %，且相比於FPID控制器加入雜訊的窗形軌跡平均誤差改善29.99 %，加入雜訊的花瓣形軌跡平均誤差改善45.13 %，證實控制器確實能有效進行音圈馬達定位控制，也使系統在具有干擾的環境下保持穩定性和強健性。The object of this study is to develop an adaptive compound fuzzy proportional-integral-derivative (ACFPID) control strategy for controlling the mover position of a voice coil motor (VCM)-based X-Y motion stage. The objectives of the study are to apply several different controllers to reduce the stability error and to compare the effects of these controllers. First, the operating principle and dynamics of the VCM-based X-Y motion stage are described. Then, a design of the fuzzy proportional-integral-derivative (FPID) control is introduced on the basis of fuzzy theory. With the additional degree of freedom to the control system parameters, the FPID control can improve the control responses and robustness of the conventional proportional-integral-derivative (PID) control. In order to improve the chattering of the system, an adaptive compound fuzzy control is proposed based on adaptive control, which can solve oscillation phenomenon of the traditional FPID. However, tuning these extra fuzzy membership function (MF) operators increases the complexity of the control system design. In this regard, the adaptive compound fuzzy feedforward controller is further proposed in which interdependent MF parameters are all online optimally determined via a Teaching-Learning-Based Optimization (TLBO). In the TLBO, the Teacher Phase and Learner Phase are dynamically adjusted to regulate the abilities of global and local searches. The summation of integral absolute errors in X-axis and Y-axis of the VCM-based X-Y motion stage during tracking process is chosen as a performance index for minimization. In this study, all of the control strategy were implemented via the digital signal processor. In addition, two reference trajectories and three different control strategies were provided to evaluate the control performances of control systems. Finally, the experimental results can verify the designed controller. When compared with the control performance of the traditional PID controller, the proposed new controller improves the average trajectory error of the window trajectory by 58.28 % after adding noise. Additionally, the average trajectory error of the flower trajectory is enhanced by 66.33 % after adding noise. In comparison to the control performance of the FPID controller, the average trajectory error of the window trajectory is improved by 29.99 % after adding noise. Similarly, the average trajectory error of the flower trajectory is enhanced by 45.13 % after adding noise. The experimental results can be verified that the designed controller can effectively control the voice coil motor positioning platform.