單參數擴張狀態觀測器之收斂性與參數選取分析

Abstract

觀測器誤差動態的收斂性與參數選取對於估測精度有關鍵性的影響，本論文中利用韓京清先生所提出的自穩定域概念，針對 單參數擴張狀態觀測器進行分析，並利用分析結果推導出使誤差狀態收斂到線性區所需之條件。之後再利用自動控制理論 對線性區的擴張狀態觀測器誤差動態進行分析並給出一參數選取法則，使得在未知函數的導函數之上界已知的情況下， 可控制估計誤差在我們所設定的範圍內。最後我們透過模擬驗證測試參數選取法則之有效性。

Parameters selection of an observer is well known to have great influence on its convergence speed and estimation accuracy. In this thesis, we will employ the self stable region concept, introduced by Dr. Han, to analyze the convergence and accuracy of a single-parameter extended state observer (ESO). A condition that guarantees the error state converging to a linear region of the ESO is proposed. Moreover, when given a desired estimation error, we also present a guideline of parameter selection to realize the demand of estimation error whenever the derivative of the unknown parameter is known. Simulation results have clearly demonstrated the effectiveness of the design.