Development of Multichannel Controller for Adaptive Optics

碩士 === 國立成功大學 === 光電科學與工程學系 === 100 === Adaptive optics system (AOS) consists of three parts: wavefront sensing, wavefront correction, and system control. The function of AOS is to compensate the wavefront aberrations induced by internal and external disturbances; hence the performance of the optica...

Full description

Bibliographic Details
Main Authors: Sih-YingChen, 陳思穎
Other Authors: Shean-Jen Chen
Format: Others
Language:en_US
Published: 2012
Online Access:http://ndltd.ncl.edu.tw/handle/80282967381456975427
Description
Summary:碩士 === 國立成功大學 === 光電科學與工程學系 === 100 === Adaptive optics system (AOS) consists of three parts: wavefront sensing, wavefront correction, and system control. The function of AOS is to compensate the wavefront aberrations induced by internal and external disturbances; hence the performance of the optical system can be improved to achieve the diffraction limit. In order not to interrupt the operation velocity as compensating the aberrations, a real-time AOS is a demand. The purpose of this thesis aims at the development of an appropriate multichannel control system which can associate our lab-made Shack-Hartmann wavefront sensor (SHWS) with the deformable mirror (DM), and then complete a real-time close-loop AOS. In the control architecture, the Zernike polynomial coefficients measured by the SHWS were fed back to create error signals, and then the optimal state feedback gain matrix based on a linear quadratic integral (LQI) control theory were derived to drive the DM for the purpose of eliminating the optical aberrations. In order to obtain a mathematical model between the SHWS and the DM, we adopt a multi-input and multi-output (MIMO) system identification method, called numerical subspace state space system identification (N4SID), to acquire a state space model according to real input and output information. With the help of this model, an off-line analysis for the system can be utilized to calculate the optimal gain parameter by the LQI theory and a simulation was tested to ensure the convergence of the control loop. At last, the matrix parameters of the control system are delivered from a personal computer to a field-programmable gate array (FPGA) module and the overall real-time close-loop AOS is implemented by the FPGA module individually.