Application of Genetic Algorithm to the Design of QFT Controller for Optical Disc Drives

• Main Author: 趙露明
• Other Authors: 姚立德
• Format: Others
• Language: zh-TW
• Published: 2001
• Online Access: http://ndltd.ncl.edu.tw/handle/30455046180144221571

碩士 === 國立臺北科技大學 === 機電整合研究所 === 89 === Quantitative Feedback Theory (QFT) is considered as an effective method for the design of robust controllers for uncertain plants to satisfy the assigned specifications. The controller design based on QFT relies heavily on designer’s expertise of controller dynamics as well as controller’s possible influence of open-loop on the gain-phase plot. The designer needs to design the QFT controller for loop-shaping by trial and error based on the gain phase plot corresponding to the specifications and variations of system dynamics. In this thesis, a methodology using the Genetic Algorithm is proposed to improve the design accuracy, to reduce the work of tedious trial and error process and to enable automatic design procedure. Methods of measuring the nominal open loop functions satisfying the bounds that transformed from the specifications on the gain-phase plot are proposed in this thesis. Examples are to be presented showing that the QFT controller designed by the Genetic Algorithm successfully satisfies the specifications. The methodology has also been applied to the design of a focus servo controller for the optical disc drive.