新型無模型反覆式學習控制演算法應用於振鏡掃描系統三角波追蹤

A Novel Model-free Iterative Learning Control Algorithm Applied to Triangular Wave Tracking on Galvanometer Scanning System

張耕豪(Geng-Hao Zhang)

指導教授 ： 陳政維

國立臺灣大學/電機資訊學院/電機工程學研究所/碩士(2020年)

https://doi.org/10.6342/NTU202002267

摘要

高精密追蹤控制對於許多應用，如半導體工業，機械製造和生物醫學產業等至關重要。以光學同調斷層掃描技術(Optical Coherence Tomography，簡稱OCT)為例，振鏡掃描系統(galvanometer scanner)用於將雷射光束反射至欲掃描位置，且振鏡系統之追蹤控制性能將直接影響輸出影像品質。為了增加影像幀率(frame rate)，需提高掃描頻率，但掃描訊號頻率可能高於振鏡系統頻寬，使系統之追蹤結果不盡理想。為了解決這個問題，基於系統模型之控制方法，如反覆式學習控制(iterative learning control, ILC)和重複控制(repetitive control, RC)經常被使用。然而，對於機電系統之建模過程通常冗長且乏味，因此在工業界並不受歡迎。為了改進系統之追蹤性能與減去系統鑑別之繁複過程，本論文提出一種新型的無模型(model-free)反覆式學習控制演算法。該演算法亦可建立系統之逆濾波器(inverse filter)。逆濾波器可用於重複控制器之設計，其中重複控制演算法對於週期訊號追蹤與週期性干擾濾除具有良好效果。本論文提出之ILC演算法已實作於商用振鏡掃描系統。無論是參考軌跡追蹤或逆濾波器之建模，我們的方法相較於其他演算法擁有更好的強健性和穩態性能。

關鍵字

反覆式學習控制 ； 無模型 ； 逆問題 ； 重複控制 ； 振鏡掃描系統

並列摘要

High-Precision tracking control is essential to various applications in semiconductor, manufacturing, and biomedical industry. For example, in optical coherence tomography (OCT), galvanometer scanners are applied to direct laser beams to the desired scanning positions. The performance of the galvanometer in tracking the scanning pattern directly affects the quality of the output images. In order to increase the image frame rate, the desired scanning pattern may contain frequency components beyond the tracking bandwidth of the galvanometer. To address this problem, model-based control scheme such as iterative learning control (ILC) and repetitive control (RC) is typically applied. However, modeling of a mechatronic system is tedious and not preferable in industry. To improve the tracking performance and to eliminate the modeling procedure, a novel model-free ILC algorithm is proposed in this thesis. The proposed model-free ILC is also extended to the application of constructing an inverse filter of the controlled system. The constructed inverse model can be applied to model-free RC design, which is excellent in tracking periodic references or rejecting periodic disturbances. The proposed model-free ILC algorithm has been implemented on a commercial galvanometer scanning system. In both reference tracking and inverse model identification, our method owns better robustness and steady-state performance than the state-of-the-art approaches.

並列關鍵字

Iterative Learning Control ； Model-free ； Inverse Problems ； Repetitive Control ； Galvanometer Scanning System

參考文獻

A. Manakov, H.-P. Seidel, and I. Ihrke, “A mathematical model and calibration procedure for galvanometric laser scanning systems.” 01 2011, pp. 207–214.

M. Wojtkowski, “High-speed optical coherence tomography: basics and applications,” Applied optics, vol. 49, no. 16, pp. D30–D61, 2010.

S. Pyshkin and J. Ballato, Optoelectronics: advanced materials and devices.

BoD–Books on Demand, 2013.

V.-F. Duma, P. Tankam, J. Huang, J. Won, and J. P. Rolland, “Optimization of galvanometer scanning for optical coherence tomography,” Applied Optics, vol. 54, no. 17, pp. 5495–5507, 2015.