撓性伺服機械系統之關節參數鑑別

Parameter Identificaton for the Joints of a Flexible Servomechanism

本研究針對多自由度撓性伺服機械系統之關節參數進行鑑別。本文延續原有敏感度最佳法之鑑別工作所推導的狀態空間方程、測試數據，利用遺傳演算法，依據目標函數與設計變數設定範圍，進行設計變數之鑑別。並針對演算程序及結果，與敏感度最佳法之鑑別比較。文中以兩個例子進行參數鑑別，範例之設計變數分別為6個阻尼參數以及6個阻尼加5個勁度參數。對系統反應參考點敏感度較大之設計變數，兩種演算法計算結果差異較小。

The objective of this research is to develop an identification method for the joint parameters in flexible servomechanisms. The derivation of dynamic and control equations, tests of frequency responses, and identification of joint parameters by a sensitivity optimization method have been conducted before. With the same test data, the parameters identification by means of genetic algorithms, according to the selected objective functions and given parameter ranges, is performed in this paper. Two examples are shown, one with six damping coefficients to be identified, while the other has five stiffness and six damping coefficients to be identified. The results are compared with those of the sensitivity optimization method. The comparison shows that the parameters with larger sensitivity have less discrepancy.