摘要 本論文主旨為在實作上以適應性基因演算法(Adaptive Genetic Algorithm,簡稱AGA)搜尋最佳化之模糊滑動控制模式 (Fuzzy Sliding Mode Controller,簡稱FSMC)參數,應用於直流無刷旋轉馬達系統,以期達到位置控制之目的。有鑒於電腦模擬之最佳控制器通常需經手調才能應用於實際系統,本論文遂以數位信號處理器建立電腦與直流無刷旋轉馬達之間的資料傳送架構,根據實際響應來進行最佳控制器之設計,並探討其性能。 本論文之控制法則的架構是以模糊控制結合滑動模式控制為主。而在控制法則之最佳化部分,分為模糊規則庫及控制法則的增益值兩部分。模糊規則庫,先使用套裝軟體Matlab/Simulink予以實現,再與控制法則的增益值同時再納入AGA即時搜尋的參數當中。實驗結果顯示,FSMC在馬達參數變動的情況下具有高強健性,而以AGA設計之FSMC可因應不同之位置命令輸入而達到最佳的輸出響應。
Abstract Investigated in this thesis is the application of an adaptive genetic algorithm (AGA) for searching optimal parameters of a fuzzy sliding mode controller (FSMC) for a DC brushless rotary motor position control system. Considering that the simulation results for optimal controllers can hardly be used directly in a real system unless through some manual adjustment, this research used a digital signal processor (DSP) as a data transmission path between the DC brushless rotary motor and a personal computer to gain real responses for controller design and performance evaluation. The structure of the control algorithm in this thesis is to combine fuzzy control with sliding mode control. In optimization for control algorithm, there are two parts of parameters. One is the fuzzy rule; the other is the controller gains. The simulation software Matlab/Simulink is to used to build up fuzzy rules, and then together with controller gains are searched by an AGA to attain their optimal values. The experimental results show that the FSMC is highly robust when the plant model has some uncertainties, and the optimal FSMC designed by AGA can track position command rapidly.