頁籤選單縮合
題名 | 整合式模糊船舶自航器可行性探討=Feasibility Study of a Unified Fuzzy Ship Steering Autopilot |
---|---|
作者姓名(中文) | 曾慶耀; 何曙光; 蘇恆平; | 書刊名 | 中國造船暨輪機工程學刊 |
卷期 | 18:4 1999.11[民88.11] |
頁次 | 頁47-54 |
分類號 | 444.6 |
關鍵詞 | 模糊控制器; 經驗語法; 數值資料對; 整合式模糊自航器; Fuzzy controller; Linguistic rule; Numerical data pair; Unified fuzzy autopilot; |
語文 | 中文(Chinese) |
中文摘要 | 本文介紹整合語言及數值兩種不同型態知識之控制方法,亦即探討結合專家經驗 法則及輸出、輸入數值資料而建整合式模糊自航器之設計方法。首先使用傳統式PD控制器 ,執行航向變換控制,並蒐集所對應之舵角、方向角及方向角速率數值資料。該資料經由模 糊化處理、區間分割、皮歸屬度演算,可建立模糊邏輯法則。每組輸出、輸入資料均可產生 一個法則,且每個法則依其價值,賦予一重要性強度指標。為避免產生才矛盾或累贅之法則 ,因此必須挑選出最高強度之法則為代表。最後則結合專家語言經驗法則,與經由處理PD 航器數值資料產生之模糊法則,而建構成整合式模糊規則庫,進而應用於自航器之設計。由 於具有專家經驗知識之領航員,及依據傳統控制理論設計之PD自航器各有其特色及優點, 而且依受控體特性不同,可能某種控制知識源較不易獲得,因此若能適當結合兩種控制知識 源以形成整合式模糊自航器,可以發揮相輔相成之效果。 |
英文摘要 | Fuzzy autopilot design method based on unification of the control knowledge from expert linguistic rules and numerical data is proposed in this paper. Specially, course-changing maneuvers is conducted via a PD type autopilot and the corresponding rudder angle, heading angle and yaw rate numerical data are collected as input-output data pairs. Fuzzy logic rules are then extracted from the collected numerical data pairs. Each input-output data pair generates a if-then rule, and a degree of importance associated with each rule is computed. Conflicting rules or redundant conditions are resolved by selecting the rule that has the highest degree of importance. The numerical data pairs- based logic rules are then combined with the expert linguistic rules in forming the unified fuzzy rule base system. Since both the human pilot and the classical control theory-based PD type autopilot are capable of steering a ship satisfactorily, the proposed method characterized by unifying the knowledge from both control systems seems to offer an attractive autopilot design method. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。