透過您的圖書館登入
IP:54.204.142.235
  • 學位論文

具殘差修正之模糊小腦模型控制器設計及其應用研究

A Study of Improving the accuracy of Fuzzy CMAC using Residual Theory Design and Its Application

指導教授 : 洪 欽 銘 許 全 守
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


傳統工業程序控制中,PID控制器使用及應用最為普遍。雖然PID控制法則簡單,但它們個別的增益參數無法隨著受控系統變化而自動調整。為了改善此一缺點,於是有著名的Ziegler-Nichols參數設定方法、模糊邏輯參數設定方法及利用生物本質為理論基礎的人工智慧控制理論等,其控制方法特色是模擬人的智能行為,不需要精確的數學模型,能夠解決許多傳統控制技術中複雜的、不確定性的、非線性的自動化控制問題。 傳統CMAC感應場中使用的是二值方盒型基礎函數,它無法儲存網路輸出入間微分之訊息,因而限制了系統參數的調整,再者,由於CMAC的近似性質及計算機造成之捨入誤差等,也因此限制了實際應用時控制精度的提昇。 因此,本論文提出將模糊歸屬函數植入傳統CMAC感應場中及新的索引指標連結規則-全連結定址架構,並結合數值分析中殘差法理論,稱為殘差修正之模糊小腦模型控制器(RFCMAC),除保有傳統小腦模型控制器之優點外,另具備輸出與系統參數之偏微分關係而得以進行參數之動態調整及控制精度的提升。 最後並將本研究所提之架構應用於線性壓電陶瓷馬達位置追隨(Tracking)控制中,藉以驗證其在實際控制系統中控制精度提昇的效能。 從實驗結果得知,無論參考模式是步階或是正弦波輸入,本研究所提具殘差修正之模糊小腦模型控制器架構在位置追隨響應上均較傳統之比例微分控制器架構更能貼近追隨於參考模式,並成功地實現本研究所提架構在線性壓電陶瓷馬達位置追隨響應上控制精度之提升。

並列摘要


For traditional industry process control, the PID controller is the most commonly used. Although the rule of PID control is simple, the main defect of PID control is that the individual gain parameters cannot be automatically adjusted when the controlled system changes. To improve this, the Ziegler-Nichols parameter setting method, the Fuzzy logic parameter setting method, and the Artificial Intelligence control theory are developed hence. The features of these methods are to imitate artificial intelligence, which can solve the problems of complexity, indetermination and nonlinear of traditional control technique, without precise mathematics model needed. The receptive field function of conventional CMAC uses the basic function of binary box, which cannot store the differential information between input and output. Also, due to the approximate property of CMAC and computer round-off error, it limits the adjustment of the system parameter and the accuracy of practical application. So, this paper presents a new index rule, which fully connects the addressing scheme and receptive field function, called RFCMAC. It not only can keep the advantages of CMAC, but also can store the differential information between input and output, which is able to auto-adjust the system parameter and improve the accuracy. Finally, to demonstrate its practical control system capability and performance of improving the accuracy, I apply the proposed structure in the position of tracking of Linear Piezoelectric Ceramic Motor (LPCM). From the experimental results, any one input of the reference model of step function or sine wave will do, the position tracking response of moving table can be closely follow the reference model compares RFCMAC with PI structure and has been successfully implemented to control the position tracking of LPCM to achieve improvement the accuracy.

並列關鍵字

PID controller Fuzzy control CMAC RFCMAC LPCM

參考文獻


[59] 張冠文,” 模糊推論積分型滑動模式之小腦模型控制器設計”,國立台灣師範大學工業教育研究所碩士論文,2002。
[45] Hahn-Ming Lee; Chih-Ming Chen and Yung-Feng Lu, “A Self-organizing HCMAC Neural Network Classifier”, Neural Networks, 2001. Proceedings , IJCNN’01.International joint Conference on ,Vol.3, pp. 1960-1965, 2001.
[1] Ziegler, Nichols,“Optimum settings for automatic controllers”, Trans. ASME, 64, 1942, pp. 759-768.
[2] L.A. Zadeh, “Fuzzy sets” , Inf. Control 8, 1965, pp.338-353.
[4] M. Brown and C.J.Harris, “The modelling abilities of the binary CMAC” , IEEE Conf. Neural Networks, vol. 3, 1994, pp. 1335 -1339.

被引用紀錄


林子華(2007)。小腦模型控制器研究〔碩士論文,國立臺北科技大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0006-1108200712310400

延伸閱讀