跳到主要內容

臺灣博碩士論文加值系統

(18.204.48.64) 您好!臺灣時間:2021/08/04 18:06
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:楊于立
研究生(外文):YANG,YU-LI
論文名稱:在水處理系統使用數據對齊實現網路攻擊識別
論文名稱(外文):Cyber-Attack Discrimination for a Water Treatment System using Data Alignment
指導教授:鄭伯炤
口試委員:李忠憲林輝堂陳煥陳嘉玫
口試日期:2020-07-30
學位類別:碩士
校院名稱:國立中正大學
系所名稱:通訊工程研究所
學門:工程學門
學類:電資工程學類
論文種類:學術論文
論文出版年:2020
畢業學年度:108
語文別:中文
論文頁數:86
中文關鍵詞:安全水處理測試平台卷積神經網路格拉姆角場網宇實體系統網路攻擊數據對齊比對
外文關鍵詞:Secure Water Treatment (SWaT)Convolution Neural Network (CNN)Gramian Angular Field (GAF)Cyber-Physical System (CPS)Cyber-attack, Data Alignment
相關次數:
  • 被引用被引用:0
  • 點閱點閱:53
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
近年來隨著工業4.0、智慧工廠的趨勢推動,現代工廠內許多自動化生產機台、設備及感測器等元件都添加上乙太網路連結,大幅增加了網宇實體系統(Cyber-Physical System, CPS)暴露於網路攻擊與駭客威脅的風險,而造成系統干擾的狀況很複雜,通常會歸因於各種來源,因此嚴重的依賴系統操作員來做出相關的對策。為了使人類決策者能夠更準確的判斷是否為網路攻擊,本文提出了一種新的方法,旨在透過網路數據與物理現象的對齊比對,以達到判斷系統遭受網路攻擊的可能性。
這項研究是在安全水處理試驗平台(SWaT)數據集上進行的,該數據集代表了現實世界中工業用水處理場的縮小版本,我們利用格拉姆角場(Gramian Angular Field, GAF)來生成圖片達成網路數據與物理現象的對齊比對,透過卷積神經網路針對產生的圖片作網路攻擊的辨別。測試數據集包括26種單級單點(Single Stage Single Point Attack, SSSP)攻擊,並發生在不同時間點及不同的階段,提出的方法經比較後結果表明,較其他的網路攻擊辨別方法有更高的精確度。

關鍵字:安全水處理測試平台、卷積神經網路、格拉姆角場、網宇實體系統、網路攻擊、數據對齊比對

With the trend of Industry 4.0 and Smart Factory, lots of automatic production machines, equipment and sensors components are connected to Ethernet, which significantly increases the exposure of cyber-physical system (CPS) to cyber-attacks and hacker threats. The conditions causing CPS disruptions are complex and usually attributed to various sources, thus manumascture operation relies heavily on system operators to make relevant countermeasures. To speed up the root cause analysis and prioritize the security incident, we propose a new method, which aims to determine the possibility of a system being attacked by cyber-attacks through the alignment of cyber data and physical phenomena.

The study was conducted on the Secure Water Treatment Testbed (SWaT) dataset which is a scaled-down version of an industrial water treatment site in the real world. We use the Gramian Angular Field (GAF) to generate images to align network data and the physical phenomena, and use the Convolution Neural Network(CNN) to discriminate the cyber-attacks with the generated images. The proposed method is more accurate than other methods for discriminating cyber-attacks.

Keywords: Secure Water Treatment (SWaT), Convolution Neural Network (CNN), Gramian Angular Field (GAF), Cyber-Physical System (CPS), Cyber-attack, Data Alignment

第一章 緒論 1
1.1 研究背景 1
1.1.1 網宇實體系統(Cyber-Physical System, CPS) 2
1.1.2 工業乙太網路(Industrial Ethernet) 4
1.1.3 入侵檢測系統(Intrusion Detection System, IDS) 10
1.1.4 機器學習(Machine Learning, ML) 13
1.2 研究動機 17
1.3 論文架構 18
第二章 相關文獻 19
2.1 機器學習解決電力系統干擾和網絡攻擊的分辨[12] 19
2.2 使用無監督機器學習作水處理系統的異常檢測[16] 22
2.3 使用卷積神經網絡檢測工業控制系統中的網絡攻擊[17] 23
2.4 相關工作比較 25
第三章 研究方法 27
3.1 概述 28
3.2 問題定義 29
3.3 系統架構 30
3.4 系統功能 31
3.4.1初階分辨器(Primary Discriminator) 36
3.4.2 格拉姆角場(Gramian Angular Field, GAF)[19] 40
3.4.3 二維卷積神經網路(2D Convolutional Neural Network) 44
第四章 實驗與結果分析 47
4.1 實驗環境說明 48
4.2 實驗用數據集 49
4.2.1攻擊場景 50
4.2.2收集過程 50
4.2.3 物理特性 52
4.2.4 網路流量性質 52
4.3 實驗情節 54
4.3.1 數據整理 54
4.3.2 格拉姆角場 58
4.3.3 應用CNN網路攻擊異常檢測[20] 61
4.4 模擬結果 66
第五章 結論與未來發展 74
參考資料 75


[1]余至浩, “關鍵基礎設施安全拉警報,專家剖析網路攻擊趨勢,” iThome, 11 5 2019. [線上]. Available: https://www.ithome.com.tw/people/130423. [存取日期: 18 12 2019].
[2]J. Lee et al., "Recent Advances and Trends in Predictive Manufacturing Systems in Big Data Environment", Manufacturing Letters, vol. 1, no. 1, pp. 38-41, 2013
[3]J. Shi, J. Wan, H. Yan and H. Suo, "A survey of cyber-physical systems", Proc. Int. Conf. Wireless Commun. Signal Process. (WCSP), pp. 1-6, Nov. 2011.
[4]Cyber physical systems: the need for new models and design paradigms. Carnegie Mellon University. p. 1–31. < https://slideplayer.com/slide/4807731/>.
[5]National Institute of Standards and Technology. Workshop report on foundations for innovation in cyber-physical systems, January 2013 <http://www.nist.gov/el/upload/CPS-WorkshopReport-1-30-13-Final.pdf/>.
[6]J. Lee and E. Lapira, "Predictive factories: the next transformation", Manufacturing Leadership Journal, vol. 20, no. 1, pp. 13-24, 2013.
[7]J. Lee, E. Lapira, S. Yang and A. Kao, "Predictive manufacturing system—Trends of next-generation production systems", IFAC Proc. Vol., vol. 46, no. 7, pp. 150-156, 2013
[8]J. Lee, B. Bagheri and H.-A. Kao, "A cyber-physical systems architecture for industry 4.0-based manufacturing systems", Manuf. Lett., vol. 3, pp. 18-23, Jan. 2015.
[9]K. Cope, “WHAT IS THE AUTOMATION PYRAMID?,” 11 6 2018. [線上]. Available: https://realpars.com/automation-pyramid/. [存取日期: 28 4 2020].
[10]Volker E. Goller, “與現有工業乙太網並存過渡 TSN翻轉工業通訊未來,” 新通訊, 04 07 2018. [線上]. Available: https://www.2cm.com.tw/2cm/zh-tw/magazine/-Technology/1FD7D00051B1466783C6A80DA5ED7795. [存取日期: 18 12 2019].
[11]Song Han, Miao Xie, Hsiao-Hwa Chen, Yun Ling: Intrusion Detection in Cyber-Physical Systems: Techniques and Challenges. IEEE Systems Journal 8(4): 1049-1059 (2014)
[12]Raymond C. Borges Hink,Justin M. Beaver,Mark A. Buckner,Tommy Morris,Uttam Adhikari,Shengyi Pan, "Machine learning for power system disturbance and cyber-attack discrimination," in 2014 7th International Symposium on Resilient Control Systems (ISRCS), 2014.
[13]Lippmann, R.: The 1999 darpa off-line intrusion detection evaluation. In: Computer Networks 34(4) 579-595, 2000. Data is available at http://www.ll.mit.edu/IST/ideval/ (2000)
[14]Bay, S.D., Kibler, D., Pazzani, M.J., Smyth, P.: The uci kdd archive of large data sets for data mining research and experimentation. ACM SIGKDD Explorations Newsletter 2(2), 81–85 (2000)
[15]Morris, T.: Industrial control system (ics) cyber attack datasets - tommy morris. https://sites.google.com/a/uah.edu/tommy-morris-uah/ics-data-set, accessed: 2016-05-06
[16]J. Inoue, Y. Yamagata, Y. Chen, C. M. Poskitt, J. Sun, "Anomaly detection for a water treatment system using unsupervised machine learning", Proc. IEEE International Conference on Data Mining Workshops (ICDMW 2017): Data 1065, 2017
[17]M. Kravchik and A. Shabtai, “Detecting cyber attacks in industrial control systems using convolutional neural networks,” in Proceedings of the 2018 Workshop on Cyber-Physical Systems Security and PrivaCy.
[18]X. Ma, Z. Dai, Z. He, J. Ma, Y. Wang, Y. Wang, "Learning traffic as images: A deep convolutional neural network for large-scale transportation network speed prediction," Sensors, vol. 17, no. 4,, p. 818, 2017.
[19]Z. Wang, T. Oates, "Encoding time series as images for visual inspection and classification using tiled convolutional neural networks", Workshops at the Twenty-Ninth AAAI Conference on Artificial Intelligence, 2015.
[20]M. Tan and Q. Le, "EfficientNet: Rethinking model scaling for convolutional neural networks", Proc. 36th Int. Conf. Mach. Learn., pp. 6105-6114, 2019.
[21]“Secure Water Treatment (SWaT),” http://itrust.sutd.edu.sg/research/ testbeds/secure-water-treatment-swat/, acc.: September 2017.
[22]Adepu, S., Mathur, A.: An investigation into the response of a water treatment system to cyber attacks. In: In Proceedings of the 17th IEEE High Assurance Systems Engineering Symposium (in Press) (2016)
[23]J. Faouzi, "pyts:A Python Package for Time Series Classification," [Online]. Available: https://pyts.readthedocs.io/en/latest/index.html. [Accessed 2020].
[24]Y. James, “[資料分析&機器學習] 第5.1講: 卷積神經網絡介紹(Convolutional Neural Network),” 25 12 2017. [線上]. Available: https://reurl.cc/qd96Yy. [存取日期: 20 6 2020].
[25]Jonathan Goh, Sridhar Adepu, Khurum Nazir Junejo, Aditya Mathur, “A Dataset to Support Research in the Design of Secure Water Treatment Systems,” 於 International Conference on Critical Information Infrastructures Security, 2017.
[26]陳澤世, “人工智慧視覺應用,” 工研院資通所, 2 8 2018. [線上]. Available: https://reurl.cc/62pMO5. [存取日期: 30 6 2020].

電子全文 電子全文(網際網路公開日期:20250827)
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top