標題: 對系統有雜訊、參數干擾和輸入限制情況下的穩定適應控制
Adaptive Control System for a Stable Plant with Input Amplitude Constraints, Bounded Disturbances and Parameter Variations
作者: 許世昌
Shih-Chang Sheu
林昇甫
Dr. Sheng-Fuu Lin
電控工程研究所
關鍵字: 適應控制、投影演算法、追蹤誤差、預估誤差、漸近輸入匹配;Adaptive Control, Projection Algorithm, Tracking Error, Prediction Error, Asymptotic Input Matching
公開日期: 1993
摘要: 在本論文中,我們研究:當外在雜訊和參數干擾存在,而且控制輸入信號 大小受限制時,對穩定最小相位系統的追蹤問題。我們提出一個含有死域 的投影演算法,並且分析這演算法的一些性質。同時,我們證明出整個閉 迴路系統是有界輸入有界輸出穩定。此外,我們也研究外在雜訊、參數干 擾和輸入訊號大小受限制對理想系統訊號的影響,同時,預估誤差、漸近 輸入匹配和追蹤誤差等也因它們的存在而不會趨近於零,但我們仍可找到 它們各別的上限。 而且我們找到一個條件,當受限制的控制輸入信號趨 近未受限制的控制輸入信號時,我們可以保證當時間充分大時追蹤誤差將 不會超過某個特定上限。最後,我們以一些模擬來證實本文中所推導的理 論的正確性。 In this thesis, we study the tracking problem for stable minimum phase system with control input amplitude constraints, external bounded disturbances and parameter variations. We present a projection algorithm with a dead-zone and analyze various properties of the algorithm. Meanwhile, we prove that the closed-loop system is BIBO stable for stable minimum phase systems. Furthermore, we explore how the disturbances, parameter variations and constraints on the control input signal influence the ideal system signals. Meantime, prediction error, asymptotic input matching and tracking error cannot approach zero when disturbances and parameter variations exist in the working environment, but we can determine upper bounds on prediction error, asymptotic input matching and tracking error respectively. And we find a condition to ensure the upper bound of tracking error when the unconstrained control input signal approach the constrained control input signal as time tends to infinity. Finally, some simulation results are given to demonstrate the validity of the theoretical results.
URI: http://140.113.39.130/cdrfb3/record/nctu/#NT820327041
http://hdl.handle.net/11536/57758
顯示於類別:畢業論文