Abstract
論文中提出一個模糊數位鎖相迴路來估測傳輸信號的相位，根據最大可能 估測法，模糊數位鎖相迴路是將傳統數位鎖相迴路加入模糊模式。它是根 據專家對相位誤差的判別，所得到的模糊規則來調整增益的大小，進而增 進收斂性能，同時可建立一個簡單的查表目錄來代替。從電腦模擬的結果 顯示，經由本論文中所提出之方法可提升收斂性能。第一章簡介，簡單介 紹傳統數位鎖相迴路之用途以及設計之考量與優缺點，接著簡單提出為何 設計模糊數位鎖相迴路之動機。第二章問題描述，首先對我們所要處理的 信號作數學描述，我們採用最大可能估測法。同時應用我們所提出的遞迴 式法來估測信號的相位，並且建立圖解來描述。第三章模糊模式，模糊模 式是結合專家經驗和數值資料，我們提出六個設計步驟來建構模糊數位鎖 相迴路。第四章收斂分析，本章針對收斂速度穩定性作分析，適當選擇步 階大小，由結果得到只要落在我們提出的區間內，可以保証演算法收斂。 第五章電腦模擬，比較模糊數位鎖相迴路和傳統數位鎖相迴路方式，電腦 模擬結果顯示模糊數位鎖相迴路的收斂速度較快，且有更好的強健性。第 六章結論，專家經驗和遞迴式演算法結合, 可增加演算法的智慧，因此模 糊數位鎖相迴路具有較快的收斂速度亦保証收斂穩定，同時可建立一個簡 單的查表目錄來代替。 In this study a fuzzy rule-based digital phase locked (DPLL) loop algorithm is proposed for phase estimation of various transmission signals. According to the maximum likelihood estimation, a fuzzy rule-based recursive algorithm is derived to obtain a DPLL to achieve phase estimation. A fuzzy rule- based method is introduced according to the expert's knowledge about phase error to adjust the loop gain and improve the performance of convergence rate. Furthermore, a simple equivalent look-up table is also developed to replace the fuzzy rule-based modeling to simplify the proposed algorithm. The simulation results have indicated that the proposed algorithm is a fast convergence scheme for phase estimation. Chapter I briefly introduces the applications of the conventional DPLL, and the motivation of designing the fuzzy ruled DPLL is stated. In chapter II, we formulate the problem and propose a new algorithm which combines a fuzzy rule model with a recursive algorithm, called a fuzzy rule-based DPLL. In chapter III, fuzzy rule model combines the expert information and numerical information, we introduce the fuzzy rule-based model and prpose a six step design procedure for generating fuzzy membership function and show how to use these fuzzy rules to obtain the proposed DPLL algorithm. In chapter IV, the analysis for convergence of the proposed DPLL algorithm is presented. The convergence of the proposed DPLL algorithm to true value can be guaranteed by choosing adequate step size of the recursive algorithm. In chapter V, the simulation results, it is seen that the proposed DPLL can track the true phase more faster than the conventional DPLL and indicate that the proposed DPLL algorithm based on fuzzy look-up table has fast convergence. In chapter VI, the simulation results have indicated that the proposed algorithm is a fast convergence scheme for phase estimation.