本論文主旨在建立一個天線信號具線上學習(On-line learning)功能之類神經網路系統。論文主題包括類神經網路演算法的推導、六種演算法的訓練比較、信號相似度比較。
我們實現的類神經網路架構是 401 - 6 - 4，401個輸入層節點、6個隱藏層節點以及4個輸出層節點，而類神經網路中所使用的權重值則利用MATLAB去做的訓練。主要整個網路設計重點在於使用最少的節點數發揮最高的運算效能，從類神經網路演算法簡化以提高天線信號運算速度。
手機通話中信號比較弱的時候或有其它鄰近頻道干擾時常會使得信號靈敏度降低，導致通話品質變差。台灣通訊業目前用到的頻帶有GSM850、GSM900、GSM1800，本論文將以減去雜訊平均值的方式提升信號相似度，藉以提高通話品質，即使在信號比較弱的地方也可以得到較好的通話品質。

The main theme of this research paper is based on the neural network system with antenna signal online learning capabilities. The thesis also discusses neural network algorithm and the comparison of six training algorithm and signal similarity.
The achieved neural network structure is 401-6-4: 401 input layer nodes, 6 hidden layer nodes, and 4 output layer nodes. MATLAB was used to run the neural network calculations. The main focus of the entire network is to use the minimum number of nodes for maximum operator efficiency, then simplify neural network algorithm to improve antenna signal processing speed.
During a mobile phone call, when the signal is weak or when there is interference from nearby channels, signal sensitivity may decrease, affecting the call quality. Currently, Taiwan has GSM850, EGSM900, and DCS1800. This paper aims to improve signal similarity by subtracting noise average, which will then increase call quality, enabling better call quality even in weak signal environments.

[1] Rosenblatt, F., “The Perceptron：A Probabilistic Model for Information Storage and Organization in the Brain,” Psych. Rev. 65, pp. 386-408, 1958.
[2] 葉怡成，類神經網路模式應用與實作，儒林圖書有限公司，2002 年。
[3] 羅華強 編著，類神經網路-MATLAB的應用，高立，中華民國94年。
[4] Rumelhart, D. E., Hinton G. E., and Williams R. J., “Parallel Distributed Processing:Explorations in the Microstructure of Cognition,” Mit Press Computational Models of Cognition and Perception Series, Vol. 1, pp. 318-362, 1986.
[5] Hagan, M. T. and M. B. Menhaj,“Training Feedforward Networks with the Marquardt Algorithm,”IEEE Transaction on Neural Networks, Vol. 5, No. 6. November 1994.
[6] Levenberg, K., “A Method for the Solution of Certain Problems in Least Squares,” Quarterly of Applied Mathematics, Vol. 2, pp. 164-168, 1944.
[7] Marquardt, D., “An Algorithm for Least Squares Estimation of Nonlinear Parameters.” SIAM Journal on Applied Mathematics, Vol. 11, pp. 431-441, 1963.
[8] Hagan,D.B.原著，類神經網路設計(Neural Network Design)，汪惠健審校，普林斯頓國際有限公司，台北，2004 年。
[9] 周鵬程 編著，類神經網路路門-活用MATLAB，全華，2006年。
[10] 張智星 著，MATLAB 程式設計與應用，清蔚科技，2000年。
[12] 梁淑芳，「導入LM 法之平行倒傳遞演算法」，東吳大學商學院資訊科學系碩士論文，2006 年。
[13] 王 志 遠，「即時手部動作辨識系統之研發」，私立中原大學醫學工程學系碩士學位論文，2002 年。
[14] 李正彬，「一個新的神經網路學習法─蟻窩演算法」，元智大學工業工程與管理學系碩士學位論文，2001 年。
[15] Yu Su Shiguang Shan, Xilin Chen Wen Gao，“Patch-Based Gabor Fisher Classifier for Face Recognition”，Pattern Recognition, 2006. ICPR 2006. 18th International Conference on Volume 2, 0-0 0 Page(s):528 – 531.
[16] Chengjun Liu; Wechsler, H.，“A Gabor feature classifier for face recognition”，Computer Vision, 2001. ICCV 2001. Proceedings. Eighth IEEE International Conference on Volume 2, 7-14 July 2001 Page(s):270 - 275 vol.2 .
[17] Rafic, A. A, and A. B. Magdy “Efficient Mapping Algorithm of Multilayer Neural Network on Torus Architecture,” IEEE Trans. on Parallel and Distributed Systems, Vol. 14, No. 9, pp. 932-943, Sep. 2003。