臺灣博碩士論文加值系統

全球顯示器產業的競爭一直處於非常激烈的狀況，台灣在競爭當中脫穎而出，創造出的產值為全球第二，目前僅次於半導體產業。本研究針對目前顯示器產業中的相關檢測方式進行強化，薄膜液晶顯示器（Thin-Film Transistor Liquid Crystal Display，TFT-LCD）在出廠前必須經過LCM Aging與LCM光學檢測。其中，LCM Aging檢測，是利用高溫的方式加速LCM模組老化，藉此檢測出LCM內的材料是否受高溫影響而損壞，另一部分為LCM光學檢測，透過色彩分析儀測量LCM的光學特性參數。

本研究建置的AIOT之LCM檢測與分析平台，利用本實驗室所開發的LCM驅動與檢測裝置來蒐集LCM Aging檢測時LCM的電壓電流資料，以及LCM光學檢測所量測到的資料，並透過所設計的CNN模型進行分析後，找出數值有異常的LCM並對產品進行分級，最後將資料傳送至所建置的雲端平台，使檢測人員可以透過遠端的方式來進行檢測，進而加快檢測速度與提高整體生產的良率；為了讓LCM驅動與檢測裝置可以驅動不同廠牌、型號的LCM，在韌體更新的部分使用FOTA的技術進行韌體更新與維護。

LCM檢測與分析平台主要實作網頁伺服器、資料庫伺服器、FTP伺服器、MQTT伺服器、LCM光學檢測CNN網路模型這五個部分，搭配自行設計的 RESTful API，即可提供LCM驅動與檢測裝置與雲端平台之間的資料交換與更新。最後，在測試結果中，透過FOTA遠端韌體更新LCM驅動與檢測裝置的韌體所花費的時間約為1分12秒，對比人工替換可省下80%以上的時間，在LCM光學檢測上透過所設計的CNN網路模型做分級預測可達到95%以上的準確率。也為了讓本研究更符合實際場域的應用，因此本研究也使用本校所建置的5G場域，來驗證整體系統的可行性與穩定性，目前也將系統移植到廠商的實際場域上做測試，希望透過實際測試來持續改善研究的不足之處。

The competition in the global display industry has always been in the cutthroat world. However, Taiwan has stood out from the competition and created the world's second largest output value, which is currently second only to the semiconductor industry. This research focuses on fortifying the current related inspection methods in the display industry. Thin-Film Transistor Liquid Crystal Display (TFT-LCD) must undergo LCM Aging and LCM Optical inspection before leaving the factory. Among them, the LCM Aging test uses high temperature to accelerate the aging of LCM module. Through this procedure, we can detect whether the materials in the LCM are damaged by the high temperature. Another part is LCM optical test, which measures the optical color properties of the LCM through a color analyzer.

Because of the research, we built the LCM detection and analysis platform. Mainly using the LCM driver and detection device to collect the current voltage of the LCM during LCM Aging detection, as well as the data measured by the LCM optical detection. Moreover, convolutional neural network model which we have designed after analysis to identify abnormal values and LCM product grading. Finally, it sends data to cloud platform in order that personnel monitor can perform the detection through the remote method. Therefore, it speeds up the detection and improve the overall production yield. In order to allow the LCM driver and detection device to apply to different brands and models of LCM, FOTA technology is used for firmware update and maintenance in the firmware update part.

The LCM detection and analysis platform mainly implements the six parts of web server, database, FTP server, MQTT server, LCM optical inspection and convolutional neural network model. With a self-designed RESTful API, it can provide LCM drivers and detection device to update and exchange data with cloud platform. Lastly, in the test results, it takes one minute and twelve seconds to update the firmware of the LCM driver and detection device through FOTA remote firmware. We can find out that it saves more than 80% of the time compared to manual replacement. In LCM optical detection the hierarchical prediction through the designed convolutional neural network model can achieve an accuracy of more than 95%. Hence, in order to make this research more in line with the actual field application, we use 5G field which was built by our school to verify the feasibility and stability of the overall system. Currently, the system is also transplanted to the actual field of the manufacturer. Furthermore, we hope that the system can continuously improve the deficiencies of the research through actual tests.

摘要...i
Abstract...ii
誌謝...iii
目錄...iv
表目錄...vi
圖目錄...vii
第一章 緒論...1
1.1 研究背景...1
1.2 動機與目的...2
1.3 國內外之相關研究情況...3
1.4 論文章節概述...4
第二章 系統原理...5
2.1 Over-the-air技術...5
2.2 Docker...7
2.3 MQTT技術...9
2.3.1 MQTT通訊格式...9
2.3.2 MQTT QoS傳輸品質...12
2.3.3 MQTT通訊原理...14
2.4 Laravel Framework...14
2.5 MD5演算法...17
2.6 5G...18
2.6.1 核心系統 (Core)...19
2.6.2 基站系統(NR)...21
第三章 系統設計...24
3.1 系統架構...24
3.1.1 LCM Aging檢測...25
3.1.2 LCM 光學檢測...26
3.2 LCM驅動與檢測裝置...30
3.3 FOTA遠端韌體更新技術...32
3.4 LCM檢測與分析平台建置與設計...33
3.5 網頁畫面與資料庫設計...37
3.5.1 Sign Up功能頁面...37
3.5.2 Sign In功能頁...37
3.5.3 LCM Aging Status功能頁面...38
3.5.4 FOTA功能頁面...39
3.5.5 Device功能頁面...40
3.5.6 LCM光學檢測功能頁面...40
3.5.7 Firmware功能頁面...41
3.5.8 LCM Model功能頁面...44
3.5.9 資料庫設計...45
3.6 5G專網...46
3.7 LCM分類分群...48
第四章 系統測試與驗證...51
4.1 LCM Aging檢測與遠端更新子系統測試...51
4.1.1 LCM Aging檢測節點通訊驗證...51
4.1.2 FOTA遠端韌體更新驗證...54
4.2 LCM光學檢測子系統測試...58
4.2.1 CNN網路模型學習結果測試...58
4.2.2 LCM光學檢測節點通訊驗證...60
4.3 安全性...62
4.3.1 LCM檢測與分析平台...62
4.3.2 MQTT協定加密...64
4.3.3 雲端平台身分授權與驗證...65
4.4 4G與5G下載測試與比較...68
第五章 結論與未來展望...72
5.1 結論...72
5.2 未來展望...73
參考文獻...74
Extended Abstract...76

[1]工業4.0，https://en.wikipedia.org/wiki/Fourth_Industrial_Revolution。
[2]朱海成，2017管理資訊系統。
[3]王泰裕，科學發展周刊第544期，工業4.0時代來臨。
[4]從工業4.0到生產力4.0，https://www.digitimes.com.tw/iot/article.asp?cat=158&id=0000437976_dqxl0bkj6wmw197a1gsnb。
[5]IoT 2020 in Review: The 10 Most Relevant IoT Developments of the Year，https://iot-analytics.com/iot-2020-in-review/。
[6]5G，https://zh.wikipedia.org/wiki/5G。
[7]State of the IoT 2020: 12 billion IoT connections, surpassing non-IoT for the first time，https://iot-analytics.com/state-of-the-iot-2020-12-billion-iot-connections-surpassing-non-iot-for-the-first-time/。
[8]行政院 (2020)，台灣顯示科技與應用行動計畫。
[9]大尺寸液晶顯示產業榮景持續至2021上半年，液晶顯示面板廠全年獲利可期， https://www.idc.com/getdoc.jsp?containerId=prAP47398421。
[10]2020 Was the Strongest Growth Year on Record for PC Monitors, According to IDC，https://www.idc.com/getdoc.jsp?containerId=prUS47545521。
[11]王建智、江行全、魏士凱 (2014)，液晶面板在高溫操作下品質客訴之改善。
[12]林鉅山，2016，”工業4.0導入產業的步驟解析”，國立中央大學高階主管企管碩士班學術論文。
[13]張嘉妤，2015，” 智慧工廠整合服務創新模式”， 東海大學高階經營管理碩士在職專班學術論文。
[14]李宜靜，2018，” 具MQTT協定之工業物聯網閘道器實現廠區製造資訊雙向互動系統”， 國立成功大學製造資訊與系統研究所學術論文。
[15]Dan Wang, Dong Chen, Bin Song, Nadra Guizani, Xiaoyan Yu, Xiaojiang Du, 2018, “From IoT to 5G I-IoT: The Next Generation IoT-Based Intelligent Algorithms and 5G Technologies”, IEEE Communications Magazine, pp. 114 - 120, November 2018.
[16]Baotong Chen, Jiafu Wan, Lei Shu, Peng Li, Mithun Mukherjee, Boxing Yin, 2017, IEEE Access, pp. 6505 - 6519, December 2017.
[17]Poonam Thakur, Varsha Bodade, Angitha Achary, Madhuri Addagatla, Neeraj Kumar, Yogesh Pingle, “Universal Firmware Upgrade Over-The-Air for IoT Devices with Security”, 2019, 2019 6th International Conference on Computing for Sustainable Global Development (INDIACom), March 2019.
[18]Neha Yadav, Dharmveer Singh Rajpoot, Shri Krishna Dhakad, “LARAVEL: A PHP Framework for E-Commerce Website”, 2019, 2019 Fifth International Conference on Image Information Processing (ICIIP), November 2019.
[19]Hans Chandra, Erwin Anggadjaja, Pranata Setya Wijaya, Edy Gunawan, “Internet of Things: Over-the-Air (OTA) firmware update in Lightweight mesh network protocol for smart urban development”, 2016, 2016 22nd Asia-Pacific Conference on Communications (APCC), August 2016.
[20]Haidi Zhu, Jingchang Huang, Huawei Liu, Qianwei Zhou, Jianqing Zhu, Baoqing Li, “Deep Learning Enabled Automatic Optical Inspection for Module-Level Defects in LCD”, IEEE Internet of Things Journal, 2021, pp. 1 - 1, May 2021.
[21]Hsueh-Ping Lu, Chao-Ton Su, “CNNs Combined With a Conditional GAN for Mura Defect Classification in TFT-LCDs”, 2021, IEEE Transactions on Semiconductor Manufacturing, pp. 25 – 33, Jan 2021.
[22]Over-the-air programming，https://en.wikipedia.org/wiki/Over-the-air_programming。
[23]空中下載技術，https://wiki.mbalib.com/zh-tw/%E7%A9%BA%E4%B8%AD%E4%B8%8B%E8%BD%BD%E6%8A%80%E6%9C%AF。
[24]Firmware Over-the-Air，https://zh.wikipedia.org/wiki/%E7%A9%BA%E4%B8%AD%E5%9B%BA%E4%BB%B6%E4%B8%8B%E8%BD%BD。
[25]ESP8266開發之旅網路篇無線更新- OTA韌體更新，https://www.itread01.com/content/1560928323.html。
[26]Docker，https://docs.docker.com/get-started/overview/。
[27]Microsoft Docs: 什麼是 Docker？，
https://docs.microsoft.com/zh-tw/dotnet/architecture/microservices/container-docker-introduction/docker-defined。
[28]何宗諭，2016，”淺談輕量化的虛擬技術-Docker容器”，國立臺灣大學計算機及資訊網路中心電子報，第0036期。
[29]Docker-從入門到實踐，https://philipzheng.gitbook.io/docker_practice/basic_concept。
[30]MQTT，https://zh.wikipedia.org/wiki/MQTT。
[31]MQTT Version 5.0，https://docs.oasis-open.org/mqtt/mqtt/v5.0/mqtt-v5.0.pdf。
[32]MQTT 通訊協定，https://openwrt-nctu.gitbook.io/project/experiment-mqtt/mqtt-protocol。
[33]MQTT教學，https://swf.com.tw/?p=1015。
[34]PHP - wiki，https://zh.wikipedia.org/wiki/PHP。
[35]PHP Documentation，https://www.php.net/manual/zh/history.php.php。
[36]Laravel，https://laravel.com/。
[37]MVC架構，https://tw.alphacamp.co/blog/mvc-model-view-controller。
[38]Github-Framework，https://github.com/topics/framework。
[39]雜湊函式，https://zh.wikipedia.org/wiki/Hash。
[40]MD5，https://zh.wikipedia.org/wiki/MD5。
[41]蕭佑和，2019，”【5G科普】只要9張圖，看懂什麼是5G”。
[42]中華電信5G，https://www.cht.com.tw/home/campaign/5g/index.html。
[43]NetCore 5，https://www.polarisnetworks.net/5g-core-network.html。
[44]3GPP R15，https://www.3gpp.org/release-15。
[45]FHES-5211，http://www.perobot.com.tw/mecindex。
[46]RPQN-7801，https://www.fhnet.com.tw/。
[47]Color Analyzer CA-310，https://sensing.konicaminolta.asia/product/color-analyzer-ca-310/。
[48]色域，https://color.viewsonic.com/zh-cn/explore/content/Color-gamut_6.html。
[49]什麼是伽瑪？，https://www.benq.com/。
[50]什麼是色溫？，https://www.benq.com/。
[51]Planckian locus，https://en.wikipedia.org/wiki/Planckian_locus。
[52]王昱翔，2021，＂LCM光學與老化檢測之物聯網系統設計＂，國立虎尾科技大學資訊工程系碩士學位論文。