針對一個未知系統進行控制，在傳統方法上必須對此未知系統進行系統識別，先求得參數化的系統的轉移函數，再進行控制器設計，此過程通常耗時複雜且識別結果不一定完全正確。數值化控制器理論（Data-Based Controller Synthesis,DBCS）提供另一種觀點，其主要精神是根據系統開迴路響應數據進行運算，因為系統響應可視為系統特性的展現，由此著手可免去繁雜的系統識別工作。本文基於DBCS理論結合根軌跡法，提出另一種演算法，即使當系統轉移函數模型未知仍可進行系統最大穩定增益值估測。此外在量測系統響應數據時，一定會有量測上的誤差存在，此誤差勢必會對估計結果有所影響。本論文亦討論各種量測雜訊等級存在時，參數估測的準確度進行模擬。

In order to control on unknown system, the most common way is proceeding the system identification. By the process, we can estimate a transform function. With the transfer function, we can design the controller. Unfortunately, the most problem is that a precise parameterized model is difficult to obtain. The property of system is imperfect as a result of the incorrect model. Data-Based Controller Synthesis (DBCS) proposes that it can avoid the above problem. DBCS is based on open-loop experiment data even though a parameterized model is unknown. By combining the root-locus method with DBCS ,this thesis develops a new algorithm to calculate the maximum stable Gain without the parameterized model. Besides, the open-loop experiment data mixed up measurement noise usually. This thesis will simulate actual data and discuss the influence of measuring noise.

[1] B.C.Kuo, ”Automatic control System”,7th ed,Prentice-Hall. 1997
[2] Chan, J. T. “ Control System Synthesis Based on Plant Test Data ”To appear in ASME/JDSMC（1994）
[3] CHARLES L. PHILLIPS ; H. TROY NAGLE 原著 陳明賢；李迪章；汪惠健 編譯 ，數位控制系統 第三版 高立圖書有限公司 民國93年。
[4] Katsuhiko Ogata , ” Discrete-Time Control System ” , Prentice-Hall Englewood Cliffs ,N.J. ,（1987）
[5] M. Frigo and S. Johnson. "Fftw: An adaptive software architecture for the fft". In ICASSP conference proceedings, volume 3, pages 1381--1384, 1998