本論文，在於應用聲學感測器來實現數位信號處理技術於各種實際對象之聲音特徵萃取。此論文選擇汽車輪胎總成(ATS)、實車引擎(AE)、工業風機(IFS)與人類(Female)說話之語音等測試數據做為演算法驗證之對象，時域與頻域等兩種演算法實現則應用適應性卡爾曼濾波器（Adaptive Kalman Filter）與梅爾頻率倒頻譜(Mel-scale frequency cepstral coefficient, MFCC)等數位濾波器。汽車輪胎總成、實車引擎與工業風機等轉動機械運轉時的聲音特徵訊號並搭配光學信號裝置，進行聲音階次能量的信號演算處理，由聲音階次能量分布可得知輪胎總成、實際引擎操作及工業風機等各使用狀況的即時特徵，用以表示轉動機械的異常特徵信號。另外，將MFCC演算法實現在人類(Female)之非週期性語音特徵解析上，可解析出身體正常與感冒狀態之差異點。
本論文採用即時演算硬體是美商國家儀器（National Instruments）的cRIO-9068嵌入式系統和NI-9234的動態信號擷取裝置進行信號擷取、即時演算分析與結果實現，可即時將機械轉動之聲音特徵經由光學脈衝信(Fiber Optical)號和麥克風聲音信號的裝置進行即時監測與分析。同時，由MFCC演算法搭配聲學信號驗證，在人類(Female)語音及機械轉動聲學信號之濾波效果。

此實驗架構與結果，將兩種演算法實現於一台三菱富利卡(Freeca)的實車輪胎總成系統，實際有效分辨輪胎的磨損即時狀態、引擎實車故障的驗證與辨識、工業風機的異常現象萃取及人類(Female)聲語音狀況的識別，兩種演算法在四種各種不同條件下的驗證結果比較，均可實現兩種演算法於實際應用系統中。

This thesis is about applying acoustic sensors to implement digital signal processing (DSP) technology to acoustic feature extractions of various practical objects. The evaluative data of automotive tire set (ATS), automotive engines (AE), industrial fan systems (IFS) and human vocals are collected and analyzed. Two analysis algorithms of time and frequency domains use adaptive Kalman filter and Mel-scale frequency cepstral coefficient (MFCC). On the other hand, by employing an MFCC algorithm, it is proved that the differences of the healthy and flu human vocal features can be distinguished. The real-time analysis system applies a National Instruments (NI) compact-RIO 9068 embedded controller and NI-9234 dynamic acquisition module for adopting to extract signals. Experimental study was carried out on the ATS of a Mitsubishi Freeca Car to effectively distinguish the real-time tire wear and engine failure confirmation. The anomalies of industrial fans and human vocal conditions were also investigated. From the study results, it is found that the proposed algorithms are able to be implemented to practical systems, which can distinguish the sound feature differences of normal and abnormal ones.

[1]
Oppenheim, A. V., Schafer, R. W., 1999, Discrete-Time Signal Processing, Prentice-Hall, Upper Saddle River, NJ.
[2]
Tafreshi, R., Ahmadi, H., Sassani, F. and Dumont, G.., 2002, “Informative wavelet algorithm in diesel engine diagnosis,” Proceedings of the 2002 IEEE International Symposium on Intelligent Control Vancouver, Canada. pp. 361 – 366.
[3]
Guo, H., Crossman, J. A. and Murphey, Y. L., 2000, “Automotive signal diagnostics using wavelets and machine learning,” IEEE Trans. on Veh. Tech., Vol. 49, pp. 1650 – 1662.
[4]
Lin, J., 2001, “Feature extraction of machine sound using wavelet and its application in fault diagnosis,” NDT&E International, Vol. 34, pp. 25 – 30.
[5]
Vold, H., and Leuridan, J., 1993, “High resolution order tracking at extreme slew rates, using Kalman filters,” SAE Paper No. 931288.
[6]
Vold, H., Mains, M. and Blough, J., 1997, “Theoretical foundations for high performance order tracking with the Vold-Kalman tracking filter,” SAE paper No. 972007.
[7]
Gade, S., Herlufsen, H., Konstantin-Hansen, H. and Wismer, N. J., 1995 “Order tracking analysis,” Brüel & Kjær Technical Review, No. 2.
[8]
Gade, S., Herlufsen, H., Konstantin-Hansen, H. and Vold, H., 1999 “Characteristics of the Vold-Kalman order tracking filter,” Brüel & Kjær Technical Review, No. 1.
[9]
Fukunaga, K., 1990, Introduction to statistical pattern recognition, Academic Press, San Diego, California.
[10]
Haji, M. and Toliyat, H. A., 2001, “Pattern recognition － a technique for induction machines rotor broken bar detection,” IEEE Trans. on Ener. Convers., Vol. 16, pp. 312 – 317.
[11]
Chih-Jung Chen, Long-Sun Chao, Tien-Hua Li, Chih-Wei Hsu, Sheng-Kai Chiu, Chih-Hui Su, Lung-Hsien Huang, Shea-Yen Gerda Eva Tai, Yu-Chieh Chuang, Cihun-Siyong Gong, Chun-Wei Su and Fu-Yang Su, ”Real Time Monitoring System for Automotive Tire Set Using an Acoustic Signal“, The 2nd International Conference on Vehicle Technology and Intelligent Transport Systems, Italy, April 23-24, 2016 (Candidate to VEHITS 2016 best student paper award).
[12]
Yu-Chieh Chuang,“Design and Implementation of Acoustic
Emission Signaling System for Online Early Fault Detection on Industrial Fans＂, Department of Electrical Engineering Chang Gung University Master Thesis, 2016。
[13]
Adaptive filter theory, 2. Simon’s Journal-JSV(J. D.Wu, M. Bai, F. C. Su, and C. W. Huang, “An expert system for the diagnosis of faults in rotating machinery using adaptive order-tracking algorithm,” Expert Syst. Appl., vol. 36, no.3, pp. 5424-5431, Apr.2009.

[14]
M. Bai, J. Huang, M. Hong,and F.C.Su, “Fault diagnosis of rotating machinery using an intelligent order tracking system,”J. Sound Vib.,vol. 280, no. 3, pp.699-718, Feb. 2005.
[15]
王小川編著，語音信號處理，全華科技圖書股份有限公司出版，2004
[16]
艾學安，＂3D聲音信號處理實現於汽車音訊:以單麥克風背景噪音估測為基礎之適應性音訊增益控制器＂，國立交通大學機械工程學系碩士論文, 2009
[17]
黃嘉銘，＂聲音及振動信號之時頻分析”，國立成功大學航空太空工程學系碩士論文，2007
[18]
Lin, C. T., 1996 Neural Fuzzy Systems. Prentice-Hall, Englewood Cliffs, NJ.
[19]
Maynard, J.D., Williams, E.G. and Lee Y., 1985, “Nearfield acoustic holography: I. Theory of generalized holography and the development of NAH,” J. Acoust. Soc. Am., Vol. 78, pp. 1395 - 1413.
[20]
Kwon, H. S. and Kim, Y. H.,1998, “Moving frame technique for planar acoustic holography”, J.Acoust.Soc.Am., Vol. 103, pp. 1734 – 1741.
[21]
Hald J., 1989, “STSF - a unique technique for scan-based Near-field Acoustic Holography without restrictions on coherence,” Brüel & Kjær Technical Review No. 1, pp. 1 - 50.
[22]
Hald, J., 2000, “Non-stationary STSF,” Brüel & Kjær Technical Review No. 1, pp. 1 - 36.
[23]
Ryan, J. G.. and Goubran, R. A., 2000, “Array optimization applied in the near field of a microphone array,” IEEE Trans. on Spee. and Aud. Process., Vol. 8, pp. 173 – 176.
[24]
Johnson, D.G. and Dudgeon, D.E., Array signal processing: concepts and techniques, Prentice Hall, New Jersey, 1993.