給定一組多訊號源經過瞬時多通道混合的量測資料，一些現存的盲蔽訊號源分離(blind source separation, BSS)演算法只能抽取出一個訊號源和估算出其相對應的通道，例如：祁忠勇博士等人提出的快速峰度最大值演算法(fast kurtosis maximization algorithm, FKMA)及渦輪式訊號源分離演算法(turbo source separation algorithm, TSSA)。然而我們的目標是要分離出所有的訊號源，所以我們必須要利用多級連續消除(multistage successive cancellation, MSC)程序來達成目的。但是多級連續消除程序會造成盲蔽訊號源分離演算法一級接續一級誤差累積的影響，而使得訊號源分離效能降低。我們將之命名為多級連續消除快速峰度最大值演算法(MSC－FKMA)及多級連續消除渦輪式訊號源分離演算法(MSC－TSSA)。在這篇論文之中，我們提出了二種新穎的多級通道限制(multistage channel-constrained, MCC)盲蔽訊號源分離演算法，分別稱為多級通道限制快速峰度最大值演算法(MCC♁FKMA)及多級通道限制渦輪式訊號源分離演算法(MCC♁TSSA)。此二種演算法是強制訊號源抽取濾波器之參數向量與前級所估測的通道參數向量互相垂直，所以能夠抽取訊號源同時又可免於誤差累積的影響。此外，本篇論文中，我們也分析了FKMA及TSSA的效能，還有證明了MCC♁FKMA及MCC♁TSSA在每一級均會抽取出不一樣的訊號源。最後，我們以一些摸擬結果來驗證此二種演算法的優異性能，和實驗真實語音及生醫訊號盲蔽訊號源分離的效能。

With a given set of multichannel measurements of instantaneous mixture of multiple sources, some blind source separation (BSS) algorithms including the fast kurtosis maximization algorithm (FKMA) and turbo source separation algorithm (TSSA) proposed by Chi et al. can only extract one source signal and the associated column of the mixing matrix . Separation of all the sources requires a multistage successive cancellation (MSC) procedure resulting in performance degradation due to error propagation effects from stage to stage. In this thesis, two novel multistage channel-constrained (MCC) BSS algorithms, referred to as MCC♁FKMA and MCC♁TSSA, are proposed which design the source extraction filter with the constraint of the source extraction filter orthogonal to all the estimated columns of obtained at all the previous stages, and the estimated source signal is free from error propagation effects at each stage. Some simulation results are presented to support that the efficacy of the proposed two novel BSS algorithms.

[1] L. Tong, V. C. Soon, Y. F. Huang, and R. Lin, “AMUSE: A new blind identification algorithm,” Proc. IEEE International Symposium on Circuits and Systems, New Orleans, LA., May 1-3, 1990, vol. 3, pp. 1784-1787.

[2] A. Belouchrani, K. Abed-Meraim, J. F. Cardoso, and E. Moulines, “A blind source separation technique using second-order statistics,” IEEE Trans. Signal Processing, vol. 45, no. 2, pp. 434-444, Feb.1997.

[3] C.-Y. Chi and C.-H. Chen, “Cumulant based inverse filter criteria for MIMO blind deconvolution: properties, algorithms, and application to DS/CDMA systems,” IEEE Trans. Siganl Processing, vol. 49, no. 47, pp. 1282-1299, July 2001.

[4] C.-Y. Chi, C.-Y. Chen, C.-H. Chen and C.-C. Feng, “Batch processing algorithm for blind equalization using higher-order statistics,” IEEE Trans. Signal Processing Magazine, vol. 20, no. 1, pp. 25-29, Jan. 2003.

[5] C.-Y. Chi, C.-H Chen and C.-Y. Chen, “Blind MAI and ISI suppression for DS/CDMA systems using HOS-based inverse filter criteria,” IEEE Trans. Signal Processing, vol. 50, no. 6, pp. 1368-1381, June 2002.

[6] C.-Y. Chi, C.-J. Chen, F.-Y. Wang and C.-H. Peng, “Turbo source separation algorithm using HOS based inverse filter criteria,” Proc. 3rd ISSPIT-03, Darmstadt, Germany, Dec. 14-17, 2003.

[7] C. Chang, Z. Ding, S.-F. Yau, and F. H. Y. Chan, “A matrix-pencil approach to blind separation of non-white sources in white noise,” Proc. ICASSP-98, Seattle, May 12-15, 1998, vol. 4, pp. 2485-2488.

[8] P. A. Delaney and D. O. Walsh, “A bibliography of higher-order spectra and cumulants,” IEEE Signal Processing Magazine, vol. 11, no. 3, pp. 61-70, July 1994.

[9] J. M. Mendel, “Tutorial on higher-order statistics (spectra) in signal processing and system theory: Theoretical results and some applications,” Proc. IEEE, vol. 79, pp. 278-305, Mar. 1991.

[10] C. L. Nikias and J. M. Mendel, “Signal processing with higher-order spectra,” IEEE Signal Processing Magazine, vol. 10, no. 3, pp. 10-37, July 1993.

[11] C. L. Nikias and A. P. Petropulu, Higher-Order Spectra Analysis: A Nonlinear Signal Processing Framework. Englewood Cliffs, New Jersey: Prentice-Hall, 1993.

[12] A. Cichocki, S. Amari, K. Siwek et al., “ICALAB for Signal Processing – benchmarks,” http://www.bsp.brain.riken.go.jp/ICALAB.

[13] Y. Inouye and K. Tanebe, “Super-exponential algorithms for multi-channel blind deconvolution,” IEEE Trans. Signal Processing, vol. 48, no. 3, pp. 881-888, Mar. 2000.