使用頻域限制時域最佳化控制器在雙軸控制平台上實現防震控制系統

Vibration Control of a Dual-axis Platform by the Frequency Constrained Time-domain Optimization Controller

紀孟江(Meng-Chiang Chi)

指導教授 ： 蕭得聖

國立交通大學/電機學院/電機學院電機與控制學程/碩士(2020年)

摘要

工業4.0帶入了許多感測器的應用整合需求，再加上大數據以及人工智慧的輔助，監控自動控制機台的即時狀態、問題反饋、機器快速的自我學習甚至是機器的自我修復，更加突顯了自動控制在實際應用上的演進。
因應現今產業的應用需求，人才的培養是一個重要的課題，本研究針對此教育需求，設計了一個雙軸控制平台系統，應用在輔助學生的教育課程上，結合理論與實作，讓學生可以學習更加完整的課程資訊，該雙軸控制平台可以透過不同的模具進行拆卸組裝，用以模擬不同控制系統，目前可模擬的控制系統如下：
1. 防震控制系統
2. 減震控制系統
3. 倒單擺控制系統
4. 阻尼控制系統
本研究主要針對上述的「防震控制系統」，將理論和實作，做一有系統的整合，模擬數位相機光學防手震的概念，主要的系統架構為上下雙層的單軸滑塊平台，下層滑塊平台可視為數位相機本體，用以模擬”手的震動訊號”，而上層滑塊平台可視為光學鏡頭。其控制目標，當下層滑塊平台移動(模擬震動訊號)，回授加速度進行位置定位控制，控制上層滑塊平台使其維持在起始的絕對位置。
本研究的實驗成果，使用QCQP方法進行控制器優化後，回授霍爾感應器的數值，上層滑塊平台可以維持在起始位置 ± 1.0mm (霍爾感應器解析度為0.109mm/Phase)，此結果和下層滑塊平台實際移動的距離無關。
若回授加速度再進行位置控制，其中因加速規的誤差、濾波器的衰減以及速度歸零機制，產生了累計誤差問題，其誤差大小結果會和下層滑塊平台的移動範圍有關，所以使用比例誤差，而實驗成果則可控制在約 4% 以內 (如. 移動100mm約產生4mm以下的誤差)。

關鍵字

防震控制 ； 最小追蹤誤差 ； 互質分解 ； Q參數化 ； QCQP ； 加速規

並列摘要

Industry 4.0 calls for integration of multiple sensors, big data, artificial intelligence, monitoring the real-time status and fault diagnosis of automatic control machines, rapid self-learning and even self-repair of machines, which highlights the evolution of automatic control in actual applications .
In response to the application demands of today's industry, cultivation of engineering professionals is an important issue. In this paper, a dual-axis control platform is built for this educational demand, which assists students in learning control theories and implementation. The dual-axis control platform can be disassembled and assembled through different modules to simulate different control systems. Currently, the control systems that can be simulated are as follows:
1. Anti-vibration control system
2. Vibration suppression control system
3. Inverted pendulum control system
4. Damping control system
This paper is aimed at the above-mentioned anti-vibration control system. It integrates control theories and implementation in a systematic way, and simulates the optical image stabilization (OIS) system of digital cameras. The system includes an upper slider and a lower slider in the same direction. The lower one can be regarded as a digital camera which suffers from “hand vibration signals". The upper one can be regarded as an optical lens. The control objective is to keep the upper slider at the initial absolute position by feeding back the acceleration of the lower slider when it vibrates.
The experimental results showed that by feeding back the position of the lower slider and optimizing the controller by the QCQP method, the upper slider can be maintained at the starting position with errors less than ± 1.0mm (the resolution of the Hall sensor is 0.109mm/Phase). This result is independent of the actual moving distance of the lower slider.
On the other hand, due to imperfection of the accelerometer, the cumulative error problem has occurred. Hence the error is related to the moving range of the lower slider. The proportional error is used here to demonstrate the control performance. The experimental results showed that the proportional error is within 4% (i.e. moving 100mm will produce an error less than 4mm).

並列關鍵字

Vibration Control ； minimum tracking error ； Coprime Factorization ； Q parameterization ； QCQP ； Accelerometer

參考文獻

[1] 王懷三, 以基因演算法則設計滑動模式控制器實現於影像感測器中二維平移式光學影像穩定器, 博士論文, 電機與控制工程研究所, 國立交通大學, 2020.

[2] C.-C. Chen, “An optical image stabilization using novel ultrasonic linear motor and fuzzy sliding-mode controller for portable digital camcorders,” IEEE Transactions on Consumer Electronics, vol. 63, no. 4, pp. 343 - 349, 2017.

[3] P. Zhao et al., “Robust Control of Large-Scale Prototypes for Miniaturized Optical Image Stabilizers with Product Variations,” in Annual Conference of the Society of Instrument and Control Engineers of Japan (SICE), Hangzhou, China, 2015.

[4] Q. Xie et al., “A Robust and Efficient Video Anti-Shaking Algorithm for Low-End Smartphone Platforms,” IEEE Transactions on Consumer Electronics, vol. 65, no. 1, pp. 1 - 10, 2019.

[5] J. Doyle, B. Francis, and A. Tannenbaum, Feedback Control Theory, pp. 63-75: Macmillan Publishing Co., 1990.