近年來氣體感測裝置逐漸被應用在各種領域，特別是在生醫領域的方面更是已有完整的應用了，而電子鼻系統就是一種應用在病患症狀監測的系統。電子鼻系統他不只整合了氣體感測裝置，更加入了隨機型機率模型分類器，如此一來在面對充滿雜訊或是資料漂移的時候更能穩定的處理資料視。在先前的文獻當中可以知道，電子鼻系統經由本身的氣體感隨機型機率模型分類器能用有效的將資料進行分辨，但隨著感測器的增加資料越來越複雜，受限於硬體限制而無法進一步分辨的效果，為了能快速的對應並減少系統的負擔，可擴展型適應性機率模型是一種新的概念，為了能解決在相同的硬體架構下能快速地進行系統資源的擴增，並同時增強處理資料的能力。
在本論文主要探討連續侷限型波茲曼演算法(Continuous restricted Boltzmann machine)於電子鼻系統的後段訊號分析與分辨的功能的應用，先使用電腦軟體MATLAB分析氣體感測資料，並提出多層結構連續侷限型波茲曼演算法，接著再由分析的結果與先前的文獻利用數位電路實現連續侷限型波茲曼演算法，接著提出兩種可擴展型的數位型連續侷限型波茲曼演算法架構，實現電子鼻系統可擴展型適應性機率模型。

In recent years, gas-sensing devices have been widely used in various fields. Especially in the field of health, medicine is already a complete application, and electronic nose system is a system used in patients with symptom monitoring. Electronic nose system, he not only integrated the gas-sensing device, but also into the random probability model classifier, to face in the face of full noise or data drift when more stable processing of information as. In the previous literature, we can see that the electronic nose system can effectively distinguish the data through its own gas-like probability model classifier. With the increase in the number of sensors more and more complex, limited by the hardware restrictions and cannot further distinguish the effect. In order to be able to quickly correspond to and reduce the burden on the system, scalable adaptive probability model is a new concept. In order to be able to solve the same hardware in the framework of the rapid expansion of system resources, and at the same time enhance the ability to process data.
In this paper, we discuss the application of the continuous restricted Boltzmann machine in the analysis and resolution of the signal after the electronic nose system. The paper analyzes the gas sensing data by using the computer software MATLAB, and proposes a multi - layer structure continuous restricted Boltzmann machine. Then the results of the analysis and the previous literature using digital circuits to achieve Continuous restricted Boltzmann machine. Then, two kinds of scalable digital continuous restricted Boltzmann machine are proposed to realize the scalable adaptive probability model of electronic nose system.

目錄
誌謝 I
摘要 II
Abstract III
目錄 IV
圖目錄 VII
表目錄 IX
第一章 緒論 1
1.1 研究動機 1
1.2 論文貢獻 2
1.3 論文內容概述 3
第二章 文獻回顧 5
2.1 電子鼻氣體資料簡介 5
2.2 連續侷限型波茲曼演算法(CRBM)簡介 6
2.2.1 連續侷限性波茲曼模型網路架構 6
2.2.2 連續侷限性波茲曼模型學習與重建模式 9
2.3 使用連續侷限性波茲曼演算法提升或降低資料維度 11
2.4 連續侷限型波茲曼演算法於積體電路之實現 13
第三章 可擴充型適應性機率模型系統架構 15
3.1 電子鼻氣體資料描述與分群效果量化 15
3.2 連續侷限型波茲曼演算法搭配後端線性分類器 18
3.2.1 分類方法 18
3.2.2 連續侷限型波茲曼演算法搭配後端線性分類器分群結果 20
3.3 具多層運算元之連續侷限型波茲曼演算法 32
3.3.1 多層連續局限型波茲曼演算法結構與學習方式 32
3.3.2 多層連續局限型波茲曼演算法的分群結果 34
3.4 以多晶片構成連續局限型波茲曼演算法 42
3.4.1 以多晶片構成大型機率模型網路 42
3.4.2 以多晶片構成多專家系統 43
3.4.3 以多晶片構成多層網路架構 44
第四章 可擴充型連續侷限型波茲曼演算法之積體電路設計 45
4.1 數位型連續侷限型波茲曼演算(DCRBM)系統設計 45
4.2 數位CRBM運作流程 51
4.3 可擴充型數位CRBM系統架構 58
4.3.1 大型機率模型網路 58
4.3.2 多專家系統 64
4.4 數位型連續侷限型波茲曼與電子鼻晶片整合 65
4.5 數位CRBM電路模擬結果 66
4.6 電子鼻晶片整合與晶片規格 71
第五章 晶片量測與結果 72
5.1 量測平台與平台設計 72
5.2 LabVIEW與資料擷取卡(DAQ)量測平台設置 74
5.3 量測結果與錯誤分析 76
5.3.1 第一版數位型連續侷限型波茲曼晶片量測結果 76
5.3.2 第二版數位型連續侷限型波茲曼晶片量測結果 78
5.4 FPGA平台驗證 80
5.4.1 FPGA開發平台介紹 80
5.4.2 FPGA量測平台設計 80
5.4.3 FPGA平台量測結果 83
第六章 結論與未來研究方向 84
6.1 結論 84
6.2 未來研究方向 85
參考文獻 86

[1] Kea-Tiong Tang, Shih-Wen Chiu, Meng-Fan Chang, Chih-Cheng Hsieh, and Jyuo-Min Shyu, “A low-power electronic nose signal-processing chip for a portable artificial olfaction system.,” IEEE transactions on biomedical circuits and systems, vol. 5, no. 4, pp. 380–90, Aug. 2011.
[2] Hung Tat Chen, Kwan Ting Ng, a Bermak, M. K. Law, and D. Martinez, “Spike latency coding in biologically inspired microelectronic nose.,” IEEE transactions on biomedical circuits and systems, vol. 5, no. 2, pp. 160–8, Apr. 2011
[3] T. J. Koickal, A. Hamilton, S. L. Tan, J. A. Covington, J. W. Gardner, and T. C. Pearce, “Analog VLSI Circuit Implementation of an Adaptive Neuromorphic Olfaction Chip,” IEEE Transactions on Circuits and Systems I: Regular Papers, vol. 54, no. 1, pp. 60–73, Jan. 2007.
[4] H. Bai and G. Shi, “Gas Sensors Based on Conducting Polymers,” Sensors, vol. 7, no. 3, pp. 267–307, Mar. 2007.
[5] N. Friedman, M. Linial, I. Nachman, and D. Pe’er, “Using Bayesian networks to analyze expression data.,” Journal of computational biology : a journal of computational molecular cell biology, vol. 7, no. 3–4, pp. 601–20, Jan. 2000.
[6] K. Strimmer and V. Moulton, “Likelihood analysis of phylogenetic networks using directed graphical models.,” Molecular biology and evolution, vol. 17, no. 6, pp. 875–81, Jun. 2000.
[7] C. C. Lu, C. C. Li, and H. Chen, “How Robust Is a Probabilistic Neural VLSI System Against Environmental Noise,” in Artificial Neural Networks in Pattern Recognition, pp. 44–53, 2008.
[8] T. B. Tang, H. Chen, and A. F. Murray, “Adaptive, integrated sensor processing to compensate for drift and uncertainty: a stochastic ‘neural’ approach,” Nanobiotechnology IEE Proc. -, vol. 151, no. 1, pp. 28–34, Feb. 2004.
[9] O. L. Mangasarian, “Linear and Nonlinear Separation of Patterns by Linear Programming,” Oper. Res., vol. 13, no. 3, pp. 444–452, Jun. 1965.
[10] Roweis, S. T.; Saul, L. K. Nonlinear Dimensionality Reduction by Locally Linear Embedding. Science. 2000, 290 (5500): 2323–2326.
[11] T. M. Cover, “Geometrical and Statistical Properties of Systems of Linear Inequalities with Applications in Pattern Recognition,” IEEE Trans. Electron. Comput., vol. EC-14, no. 3, pp. 326–334, Jun. 1965.
[12] K. G. Murty, “Linear programming,” New York: John Wiley & Sons Inc., 1983.
[13] Z. Xu, J. Chen, L. Xu, W. Wang, and A. E. Methods, “A parallel digital hardware generator of White Gaussian noise stressing on crest factor,” in 2012 18th Asia-Pacific Conference on Communications (APCC), 2012, pp. 456–461.
[14] L. Colavito and D. Silage, “Composite Look-Up Table Gaussian Pseudo-Random Number Generator,” 2009 International Conference on Reconfigurable Computing and FPGAs, pp. 314–319, Dec. 2009.
[15] W. Hörmann and J. Leydold, “Continuous random variate generation by fast numerical inversion,” ACM Transactions on Modeling and Computer Simulation, vol. 13, no. 4, pp. 347–362, Oct. 2003.
[16] G. E. P. Box and M. E. Muller, “A Note on the Generation of Random Normal Deviates,” The Annals of Mathematical Statistics, vol. 29, no. 2, pp. 610–611, Jun. 1958.
[17] W. Luk, J. D. Villasenor, and P. H. W. Leong, “A hardware Gaussian noise generator using the Wallace method,” IEEE Transactions on Very Large Scale Integration (VLSI) Systems, vol. 13, no. 8, pp. 911–920, Aug. 2005.
[18] C. Lin and J. Wang, “A digital circuit design of hyperbolic tangent sigmoid function for neural networks,” 2008 IEEE International Symposium on Circuits and Systems, pp. 856–859, May 2008.
[19] A. Mishra and K. Raj, “Implementation of a digital neuron with nonlinear activation function using piecewise linear approximation technique,” in 2007 Internatonal Conference on Microelectronics, 2007, no. December, pp. 69–72.