透過您的圖書館登入
IP:18.118.200.197
  • 學位論文

小腦模型控制器研究

A Study on Cerebellar Model Articulation Controller

指導教授 : 王順源
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


在控制系統中使用歷史最為悠久而且至今還常常在工業上使用的控制器為比例—積分—微分(PID)控制器,其原因是因為PID控制器使用方便及構造簡單。但是,PID控制器並不是每種系統都能控制的,如果將PID控制器用於一些較複雜的受控系統,或無法求出受控體的數學模型的系統,則控制性能會很差或無法控制。另外,若受控體參數有受到外在的因素影響變化時,PID控制器的參數亦無法馬上在線上調整,使其強健性能受到影響。 近年來控制的理論是朝向智慧型控制器來發展,比較有名的包含類神經網路、模糊控制、小腦模型等。小腦模型控制器 (Cerebellar Model Articulation Controller, CMAC),為類神經網路的一種,但是它具有比類神經網路計算量小且速度快的優點、結構簡單和具有線上調整參數的功能。 有鑒於PID控制器缺點,本論文針對小腦模型控制器作改善,以實現具即時調整參數能力之小腦模型PID控制器。所研究結果使用C++程式語言撰寫線上小腦模型演算法,配合Matlab/Simulink軟體模擬具有穩態誤差的二階延遲及沒有延遲系統及二階不穩定延遲及沒有延遲系統等。由模擬結果可知,小腦模型PID控制器能有效控制上列系統,且在性能上有顯著之改善。

並列摘要


The PID controller is the longest standing and the most popular controller that has ever been used in the industry because it is convenient and has a simple structure. However, the PID controller can not be used for every system. If the PID controller is used in a more complicated system or in a system where it is not possible to estimate the mathematical model of the plant, then the control performance could be very poor and the plant may not be able to be controlled. In addition, if the parameters of the plant are changed due to external factors, the parameters for the PID controller cannot be immediately modified online, thus the robustness of the PID controller may be deteriorated. In recent years, the development of control theories shows a trend towards intelligent controllers. Among them, artificial neural network (NN), fuzzy controls, and cerebellar model articulation controllers (CMAC), etc., are the most popular. CMAC is a branch of traditional artificial neural networks, and has advantages over traditional artificial neural networks, with less computation required, high computation speed, intelligent learning capability, simple structure, and the ability to adjust the parameters on-line. In view of the disadvantages of PID controllers, the CMAC PID controller, which allows real-time parameter adjustment, is studied specifically and implemented in this thesis. In this study, the C++ programming language is used for the implementation of the CMAC algorithm. Meanwhile, Matlab/Simulink is used as well to simulate the following systems: the second-order systems with steady-state errors and with/without delay, and the second-order unstable systems with/without delay. As the simulation result shows, the CMAC PID controller can effectively control the above systems and the performance is improved significantly.

並列關鍵字

CMAC PID controller

參考文獻


[12] 陳耀生,小腦模型學習速度之研究,碩士論文,中原大學電機工程研究所,民國九十二年七月。
[17] 駱佑瑋,輪型足球機器人之整合型小腦模型控制器設計與路徑規劃,碩士論文,淡江大學電機工程學系,民國九十四年六月。
[18] 王柏涵,適應性小波小腦模型之研究及其在非線性控制與影像壓縮之應用,碩士論文,元智大學電機工程學系,民國九十四年六月。
[19] 游章充,可微分小腦模型控制器於雷射二極體之溫控研究,碩士論文,臺灣大學機械工程學研究所,民國九十四年六月。
[20] 曾俊翰,模糊小腦模型控制器於高性能感應馬達控制系統之設計,碩士論文,國立臺北科技大學機電整合研究所,民國九十四年六月。

被引用紀錄


謝昀霖(2014)。班群教室室內聲環境分析與改善設計之研究〔碩士論文,國立屏東科技大學〕。華藝線上圖書館。https://doi.org/10.6346/NPUST.2014.00025

延伸閱讀


國際替代計量