金屬氧化物氣體感測器由於價格低廉，靈敏度高，穩定性良好而被廣泛用於氣體濃度預估和辨識。為了擁有良好的氣體濃度估計表現，從氣體曲線萃取穩健訊息特徵參數中所使用的方法，是氣體資料處理中非常重要的一部分。然而，對於快速達到辨識氣體濃度，從氣體曲線的暫態部分中萃取特徵參數是必要的。
在本研究中，提出了一種新的特徵萃取方法來預估氣體濃度與減少估計濃度的時間。利用一階系統(First-Order System)來擬合氣體波型資料，模型係數特徵時間常數來定義每個波型反應的速度快慢，並結合暫態特徵的特性，來達到準確的快速氣體濃度辨識。
根據氣體感測器的還原氣體反應速率，研究利用多元線性回歸來估計氣體濃度。同時對於不同的特徵萃取方法進行了氣體濃度預估比較。結果顯示，利用研究提出的暫態特徵萃取方法，其氣體濃度預估誤差小於3.62％，相較於傳統的方法，平均完成時間則可快上19倍。

Metal–oxide–semiconductor gas sensors have been widely used for gas concentration estimation and gas identification because of their low price, high sensitivity, and high robustness. To achieve optimal estimation performance, a robust feature extraction method for extracting required information from a gas response curve is vital in data processing. However, the feature extracted from the transient part of gas response is required for fast gas concentration estimations.
In this paper, we propose a new transient feature extraction method for estimating gas concentrations and reducing the estimation time. A time constant feature in a first-order system was fitted to gas response data and used to determine the response rate of each wave; combining the time constant feature with the transient feature resulted in accurate and rapid gas concentration identification.
According to the gas reaction rate, multiple linear regression was utilized for gas concentration estimation. For comparison, different feature extraction methods were investigated for gas concentration estimation. The results revealed that when the proposed approach was employed, the gas concentration estimation error was less than 3.62% and the average completion time of gas concentration estimation was reduced by 19 times relative to the times of the existing feature extraction methods.

摘 要 i
ABSTRACT ii
致 謝 iii
目 錄 iv
圖 目 錄 vi
表 目 錄 ix
第一章 緒論 1
1.1 研究背景 1
1.2 研究動機與目的 2
1.3 章節簡介 4
第二章 文獻回顧 5
2.1 氣體感測器簡介 5
2.2 特徵萃取 8
2.2.1從氣體曲線中萃取特徵 8
2.2.2 從擬合曲線中萃取參數特徵 12
2.3 多元線性回歸(Multiple Linear Regression) 13
第三章 實驗系統架構 14
3.1 訊號讀取系統 15
3.1.1 氣體感測器 15
3.1.2 氣體感測器介面電路 17
3.1.3 訊號擷取裝置 18
3.2 實驗環境控制系統 19
3.2.1 氣體產生及控制系統 19
3.2.2 氣體濃度之計算 23
第四章 氣體濃度預測 24
4.1 特徵萃取之時間常數 25
4.1.1 氣體感測器原理 25
4.1.2 特徵萃取 27
4.1.3 特徵時間常數取法 28
4.2 線性回歸預測 32
4.3 時間常數之酒精氣體濃度預測分析 33
4.3.1 單顆感測器之預測 35
4.3.2 多顆感測器之預測 37
4.3.3 Leave-One-Out Cross Validation交叉驗證之濃度預測 40
4.4 電阻變化率最大值之酒精氣體濃度預測分析 42
4.4.1 單顆感測器之預測 43
4.4.2 多顆感測器之預測 45
4.4.3 Leave-One-Out Cross Validation交叉驗證之濃度預測 47
4.5 斜率最大值之酒精氣體濃度預測分析 49
4.5.1 單顆感測器之預測 50
4.5.2 多顆感測器之預測 52
4.5.3 Leave-One-Out Cross Validation交叉驗證之濃度預測 54
4.6 結果討論 56
第五章 結論與未來展望 59
參考文獻 60

[1] L. Buck and R. Axel, “A Novel Multigene Family May Encode Odorant Receptors: A Molecular Basis for Odor Recognition,” Cell, vol. 65, pp. 175-187, Apr. 1991.
[2] Flammini A. and S. De Vito, Wireless Chemical Sensors, in Chemical Sensors Comprehensive Sensor Technologies: Volume 6 Chemical Sensors Applications edited by G. Korotchenkov, (2011), Momentum Press, ISBN: 9781606502396
[3] B. Tudu, A. Metla, B. Das, N. Bhattacharyya, A. Jana, D. Ghosh, and D. Bandyopadhyay, “Towards versatile electronic nose pattern classifier for black tea quality evaluation: An incremental fuzzy approach,” IEEE Trans. Instrum. Meas., vol. 58, no. 9, pp. 3069–3078, Sep. 2009.
[4] I. Concina, M. Falasconi, and V. Sberveglieri, “Electronic Noses as Flexible Tools to Assess Food Quality and Safety: Should We Trust Them?, ” Sensors Journal, IEEE, vol. 12, pp. 3232-3237, 2012.
[5] C. Arnold, M. Harms, and J. Goschnick, "Air quality monitoring and fire detection with the Karlsruhe electronic micronose KAMINA," Sensors Journal, IEEE, vol. 2, pp. 179-188, 2002.
[6] Y. Da-Jeng, "A gas sensing system for indoor air quality control and polluted environmental monitoring," in Nanotechnology, 2009. IEEE-NANO 2009. 9th IEEE Conference on, 2009, pp. 806-811.
[7] M. A. Ryan, K. S. Manatt, S. Gluck, A. V. Shevade, A. K. Kisor, H. Zhou, et al., "The JPL electronic nose: Monitoring air in the U.S. Lab on the International Space Station," in Sensors, 2010 IEEE, 2010, pp. 1242-1247.
[8] K. I. Arshak, C. Cunniffe, E. G. Moore, and L. M. Cavanagh, "Custom electronic nose with potential homeland security applications," in Sensors Applications Symposium, 2006. Proceedings of the 2006 IEEE, 2006, pp. 30-35.
[9] G. Dongmin, D. Zhang, L. Naimin, D. Zhang, and Y. Jianhua, “A Novel Breath Analysis System Based on Electronic Olfaction,” Biomedical Engineering, IEEE Transactions on, vol. 57, pp. 2753-2763, 2010.
[10] A. D’Amico, C. Di Natale, R. Paolesse, A. Macagnano, E. Martinelli, G. Pennazza, M. Santonico, M. Bernabei, C. Roscioni, G. Galluccio, R. Bono, E. Finazzi Agrò, and S. Rullo, “Olfactory systems for medical applications,” Sen. Actuators B, Chem., vol. 130, pp. 458–465, 2008.
[11] A. H. Abdullah, A. H. Adom, A. Y. M. Shakaff, M. N. Ahmad, A. Zakaria, F. S. A. Saad, et al., “Hand-Held Electronic Nose Sensor Selection System for Basal Stamp Rot (BSR) Disease Detection, ” in Intelligent Systems, Modelling and Simulation (ISMS), 2012 Third International Conference on, 2012, pp. 737-742.
[12] Tarik Jasarevic. 7 million premature deaths annually linked to air pollution. http://www.who.int/mediacentre/news/releases/2014/air-pollution/en/ Accessed: 2018-01-30.
[13] Egashira, M., Sessyoku-nensyo-shiki Sensa. Catalytic Combustion Gas Sensors in Kagaku Sensa Jitsuyo Benran Practical Handbook of Chemical Sensors. Fuji Tekunoshisutemu.1986, 61.
[14] Liu, X.; Cheng, S.; Liu, H.; Hu, S.; Zhang, D.; Ning, H. A Survey on Gas Sensing Technology. Sensors. 2012, 12, 9635-9665.
[15] Baker, A.R., Brit. Pat. 892, 530.
[16] Seiyama, T., et al. A new detector for gaseous components using semiconductive thin films. Anal. Chem. 1962, 34, 1502.
[17] Taguchi, N. Japanese Pat. S45-38200 (appl. For 1962).
[18] Neri, G. First Fifty Years of Chemoresistive Gas Sensors. Chemosensors. 2015, 3, 1-20.
[19] Publicity material. Gas Alarm Industries Assoc. of Japan. 1990.
[20] Ohnishi, M.; Ishibashi, T.; Kijima, Y.; Ishimoto C.; Seto, J. A molecular recognition system for odorants incorporating biomimetic gas-sensitive devices using Langmuir-Blodgett films. Sens. Mater. 1992, 4, 53–60.
[21] Balasubramanian, S.; Panigrahi, S.; Logue, C.; Gub, H.; Marchello, M. Neural networks-integrated metal oxide-based artificial olfactory system for meat spoilage identification. J. Food Eng. 2009, 91, 91–98.
[22] Distantea, C.; Leo, M.; Sicilianoa, P.; Persaud, K. On the study of feature extraction methods for an electronic nose. Sens. Actuators B Chem. 2002, 87, 274–288.
[23] Roussel, S.; Forsberg, G.; Steinmetz, V.; Grenier, P.; Maurel B.V. Optimisation of Electronic Nose Measurements. Part I: Methodology of Output Feature Selection. J. Food Eng. 1998, 37, 207–222.
[24] T. Eklov, P. Martensson, and I. Lundstrom, “Enhanced selectivity of MOSFET gas sensors by systematical analysis of transient parameters,” Anal. Chim. Acta, vol. 353, pp. 291-300, 1997.
[25] Zhang, S.; Xie, C.; Zeng, D.; Zhang, Q.; Li, H.; Bi, Z. A feature extraction method and a sampling system for fast recognition of flammable liquids with a portable E-nose. Sens. Actuators B Chem. 2007, 124, 437–443.
[26] FIGARO USA., INC. http://www.figarosensor.com/
[27] The MatlabWorks, INC. https://www.mathworks.com/products/matlab.html