近年來，航空器失事調查的方向已不再侷限於利用飛航資料紀錄器(FDR)來找尋可能的失事原因;座艙通話紀錄器(CVR)亦然成為另一個提供失事調查的有利工具。本論文利用未知訊號分離(BSS)的方法來分離座艙通話紀錄器所錄製的聲音，以提高聲音辨識率。文中以實際錄製於ATR-72民航機座艙內的聲音為訊號來源，以驗證此分離方法；而利用了類神經網路(BP)來辨識聲音訊號，訊號資料庫的來源亦為實際錄製於民航機中的座艙。本文所提出的聲音訊號分離證實可以有效的提高辨識率，且聲音訊號分離與辨識的技術證實可以有效提供飛航失事調查與監視航空器之健康資訊；本方法亦提供找尋可能損壞之元件等飛航安全應用上之方法及工具。

In recent years, the investigation of aviation accident have no longer limited to the Flight Data Recorder (FDR) to find out the possible causes of the accident. In steady, the Cockpit Voice Recorder (CVR) is also becoming another powerful too for the investigation of the accident. This thesis hence focus on the sound separation of the CVR data by using the Blind Signal Separation (BSS) technique in this thesis. Sound identification is another point in this research. By using the Back Propagation Neural Network (BP), the sound could be identified. Sound separation and identification could be contributed to the investigation of the aviation accident and the maintenance of the aircraft.

[1] ASC : http://www.asc.gov.tw
[2] NTSB : http://www.ntsb.gov/Aviation/CVR_FDR.htm
[3] Sendzimir, V., “Block Box”, Invention and Technology, Fall 1996, pp.26-32.
[4] ALPA : http://www.acoustics.org/press/133rd/2psal.html
[5] Stearman R.O, Schulze G.H., Rohre S.M., “Aircraft Damage Detection from Acoustic and Noise Impressed Signals Found by a Cockpit Voice Recorder”, Institute of Noise Control Engineering of the USA,1997
[6] Stearman R.O, Schulze G.H., and Rohre S.M., Buschow M.C., “Aircraft Damage Detection from Acoustic Signals Found by a Cockpit Voice Recorder”, Acoustical Society of America 133rd Meeting Lay Language Papers, June 17, 1997.
[7] www.rock.lz.gs.cninfo.net
www.leiwen.tripod.com
www.hull.ac.uk/music
[8] O. M. Mracek Mitchell, Carolyn A. Ross, and G. H. Yates. “Signal processing for a cocktail party effect.”, The Journal of the Acoustical Society of America, May 1971
[9] Paris J. Smaragdis, “Information Theoretic Approaches to Source Separation”, Master Thesis, Massachusetts Institute of Technology, School of Architecture and Planning, Media Arts and Sciences, 1997

[10] D. Yelling and E. Weinstein, “Criteria for multi-channel separation”. IEEE Trans. Signal Processing,42, 1994
[11] Aapo Hyvarinen, Juha Karhunen, Erkki Oja, “Independent Component Analysis”, A volume in the Wiley Series on Adaptive and Learning Systems for Signal Processing, Communications, and Control Simon Haykin, Series Editor.
[12] P. Comon, “Independent component analysis, a new concept?” Signal Processing, 36, 1994
[13] C. Jutten and J. Herault, “Independent components analysis versus principal components analysis”, In Signal Processing, volume IV of Theories and Applications, Elsevier, Amsterdam, 1988.
[14] C. Jutten and J. Herault, “Blind separation of sources, part I : an adaptive algorithm based on neruomimetic architecture.”, Signal Processing, 24, July 1991.
[15] Dominic C.B. Chan, “Blind Signal Separation”, Ph.D Thesis, University of Cambridge, Department of Engineering, Signal Processing and Communication Laboratory.
[16] http: //www. Dalterio.demon.co.uk/projects/nnn/html/contents.html
[17] Arbib, M.A., “Brains, Machines, and Mathematics”, 2nd edition, New York: Springer-Verlag (1987)
[18] Simon Haykin, “Neural Networks ~ A Comprehensive Foundation”, Prentice Hall, 1994
[19] Anderson, J.A, “General introduction”, In Neurocomputing: Foundations of Research, Cambridge, MA: MIT Press.
[20] Mendel, J. M., and R. W. McLaren, “Reinforcement-learning control and pattern recognition systems.”, New York: Academic Press.
[21] Churchland, P.S., and T.J. Sejnowski, “The Computation Brain”, Cambridge, MA: MIT Press.