臺灣博碩士論文加值系統

本文提出以FPGA來實現自建型模糊類神經網路（self-constructing fuzzy neural network, SCFNN）模糊神經網路部分，並且應用在數位通道等化器。SCFNN是一個可即時學習後，自行建立系統的模糊類神經網路，以經過學習的SCFNN擔任數位通道等化器可將通道失真信號有效的還原。在硬體實現方面，我們針對原始的演算法加以改良，再利用多工的技術在電路上節省了10M個DSP blocks與920M個logic elements，其中M 代表 SCFNN 架構中第三層的節點數目，大量地減少硬體成本與運算複雜度，以符合實際設計上的考量。

The FPGA realization of a self-constructing fuzzy neural network (SCFNN) and its applications to digital channel equalizer are presented in this thesis. SCFNN is capable of learning on-line and constructing the neural fuzzy network system structure by itself. Applying the well-trained SCFNN in a digital channel equalizer can effectively restore the distorted signals. By modifying the original algorithm and introducing multiplexing techniques in the design of the circuit, we have reduced the hardware cost of the equalizer to 10M DSP blocks and 1201M logic elements, where M denotes the number of nodes in the third layer of the SCFNN structure. The lower computational complexity of the system fits well to practical design considerations.

摘 要 i
ABSTRACT ii
誌 謝 iii
目錄 iv
表目錄 vi
圖目錄 vii
一、緒論 1
1.1 研究動機 1
1.2 研究目的 2
1.3 研究方法與內容 3
1.4 各章提要 3
二、傳輸通道及等化器 4
2.1 簡介 4
2.1.1 多重路徑傳輸通道簡介 4
2.1.2 ISI效應 5
2.2 系統架構 7
2.3 複數非線性通道 8
2.4 適應性等化 8
2.4.1 假雜訊序列(pseudo-noise sequences, PN sequences) 9
2.4.2 最小均方演算法(least mean square algorithm, LMS) 10
2.4.3 等化器之操作 12
三、類神經網路簡介 14
3.1 人腦細胞組織 15
3.2 類神經網路簡介 16
3.3 類神經網路如何運作 19
3.4 類神經網路的特色 20
四、自建式模糊類神經網路 21
4.1 SCFNN結構與推論輸出 21
4.2 結構學習演算法 22
4.3 參數學習演算法 23
五、硬體實現 25
5.1 系統整體架構 25
5.2 SCFNN硬體實現架構圖與演算法 28
5.2.1 推論輸出部份 28
5.2.2 自建結構部份 32
5.2.3 參數更新部份 33
5.2.4 多工共用 37
5.3 硬體效能評估與分析 39
5.3.1 位元錯誤率效能比較 40
5.3.2 學習曲線 41
5.3.3 硬體成本分析 44
六、結論 46
參考文獻 47
簡 歷 49

[1]M. Ibnkahla, “Applications of neural networks to digital communications,” Signal Processing,pp. 1185-1215, 80(2000).
[2]S. Chen, G. Gibson, C. Cowan, and P. Grant, “Adaptive equalization of finite non-linear channels using multi-layer perceptron,” Signal Processing, pp.107-119, 20(1990).
[3]S. Chen, B. Mulgrew, and P. M. Grant, “A clustering technique for digital communication channel equalization,” IEEE Trans. on Signal Processing, pp. 107-119, 20(1990).
[4]G. Kechriotis, E. Zervas, and E. S. Manolakos, “Using recurrent neural networks for adaptive communication channel equalization,” IEEE Trans. on Neural Networks, vol. 5, pp. 267-278, Mar. 1994.
[5]J. C. Patra, R. N. Pal, R. Baliarsinghn and G. Panda, “Nonlinear channel equalization for QAM signal constellation using artificial neural networks,” IEEE Trans. on System, Man, Cybernetics, vol. 29, no. 2, Apr. 1999.
[6]S. Ong, S. Choi, C. You and D. Hong, “A decision feedback recurrent equalizer for digital communication,” IEEE Trans. on magnetic, vol. 33, pp. 2967-2969, Sep. 1997.
[7]S. Horokawa, T. Furnhashi, and Y. Uchikawa, “On fuzzy modeling using neural networks with the backpropagation algorithm,” IEEE Trans. on Neural Networks, vol. 3, pp. 801-806, Sep. 1992.
[8]Y. Lin and G. A. Cunningham, “A new approach to fuzzy-neural system modeling, IEEE Trans. on Fuzzy System,” vol. 3, pp. 190-197, May 1995.
[9]B. Kosko, Neural networks and fuzzy systems, Englewood Cliffs, NJ: Prentice-Hall, 1992.
[10]C. T. Lin, Neural fuzzy control systems with structure and parameter learning. New York: World Scientific, 1994.
[11]I. Cha and S. Kassam, “Channel equalization using adaptive complex radial basis function networks,” IEEE J. Select. Areas Commun., vol. 13, pp122-131, Jan. 1995.
[12]F. J. Lin, C. H. Lin, and P. H. Shen, “Self-constructing fuzzy neural network speed controller for permanent-magnetic synchronous motor drive,” IEEE Trans. on Fuzzy System, vol. 9, no.5, pp.751-759, Oct. 2001.
[13]J. Shing and R. Jang, “ANFIS: Adaptive-network-based fuzzy inference system,” IEEE Trans. on Systems, Man, and Cybernetics, vol. 23, no.3, pp. 665-685, May/Jun. 1993.
[14]Altug S., Chow M. Y. and Trussell H. J., “Fuzzy inference system implemented on neural network architectures for motor fault detection and diagnosis,” IEEE Trans. on Industrial Electronics, vol. 46, no. 3, pp.1069-1079, 1999.
[15]Djukanovic M. B., Calovic M.S., Vesovic B. V. and Sobajic D. J., “Neuro-fuzzy controller of low head hydropower plants using adaptive-network based fuzzy inference,” IEEE Trans. on Energy Conversion, vol. 12, no.4, pp. 375-381, 1997.
[16]翁芳標，民國87年， “通訊工程概論”，台北市，全華科技圖書股份有限公司
[17]李宗霖，民國88年，“DAB數位音訊廣播系統之COFDM技術簡介”，無線電界，11月
[18]S. Chen, S. Mclaughlin, and B. Mulgrew, “Complex-valued radial basis function network, Part II: Application to digital communications channel equalization,” Signal Processing, vol. 36, pp. 165–188, Feb. 1994
[19]S. Haykin, Communication Systems, 4th ed., New York: Wiley, 2001.
[20]羅華強，民國90年，“類神經網路-MATALB的應用”，新竹市，清蔚科技股份有限公司