在許多不同工程領域中處理訊號問題時，時常會操作在擁有固定干擾的衰減通道模型下。故本篇論文在此環境下，提出了一個演算法來量測觀察訊號中的干擾大小、通道係數及雜訊能量。此演算法是基於最大概似估測法(maximum-likelihood estimation)，先求出干擾大小、通道係數及雜訊能量的估測器，由於以上估測器皆無封閉式的解(closed form solution)，故使用近似線性估測器解得初始值，再利用迭代法逼近最佳解。為了效能比較，我們亦求得牛頓-瑞夫生法(Newton-Raphson method)最大概似估測。根據模擬結果，我們提出的概似函數(likelihood function)具有全域最大值且收斂是被保證的。模擬結果也顯示了，我們提出的近似線性演算法具有快速收斂的能力，以及在不同情況下，都比傳統方法有較佳的效能。

The model of memoryless Gaussian channel with deterministic interference is commonly adopted in various engineering applications. In the thesis, we propose an algorithm for the measurement of channel coefficient, interference level, and noise power by observed signals over the memoryless Gaussian channel. This algorithm applies an iterative approach based on maximum-likelihood estimation for channel coefficient, interference level, and noise power. However, there isn’t a closed-form solution for these estimators. So the proposed algorithm applies an iterative approach with the initial point determined by the linear approximation of the hyperbolic tangent function. In the thesis we also use Newton-Raphson method to obtain another maximum-likelihood estimation for comparison. According to simulations, the proposed likelihood function has a global maximum and its convergence is guaranteed. Simulation results also demonstrate that, under many different situations, the proposed approach has faster convergence speed and better performance than closed-form approximation estimation.

[1] B. M. Oliver, “Thermal and Quantum Noise,” Proc. IEEE, vol. 53, no. 5, pp. 436-454, May 1965.
[2] T. S. Rappaport, Wireless Communications – Principles and Practice, Prentice Hall PTR, 2nd edition, 2002.
[3] J. G. Proakis and M. Salehi, Digital communications, McGraw-Hill, 5th edition, 2008.
[4] B. Sklar, “Rayleigh fading channels in mobile digital communication systems part I: characterization,” IEEE Commun. Mag., vol. 35, issue 9, pp. 136–146, Sep. 1997.
[5] G. L. Stüber, Principles of mobile communication. Kluwer Academic Publishers, 2nd edition, 2002.
[6] F. Chen,“Maximum likelihood based measurement of interference level and noise power for memoryless Gaussian channel with deterministic interference,”IEEE Trans. Commun., vol. 60, no. 1, pp. 19-22, Jan. 2012.
[7] N. Merhav, “Universal decoding for memoryless Gaussian channels with a deterministic interference,” IEEE Trans. Inf. Theory, vol. 39, no. 4,pp. 1261–1269, July 1993.
[8] S. H. Chung, V. Krishnamurthy, and J. B. Moore, “Adaptive processing techniques based on hidden Markov models for characterizing very small channel currents buried in noise and deterministic interferences,” Philosophical Trans.: Biological Sciences, vol. 334, no. 1271, pp. 357–384, Dec. 1991.
[9] W. J. J. Roberts and Y. Ephraim, “An EM algorithm for ion-channel current estimation,” IEEE Trans. Signal Process., vol. 56, no. 1, pp. 26–33, Jan. 2008.
[10] Y. Chen, “NDA estimation of SINR for QAM signals,” IEEE Commun. Lett., vol. 9, no. 8, pp. 688–690, Aug. 2005.

[11] H. Xu, Q. Huang, and H. Zheng, “The NDA SNR estimation for QAM signals,” in Proc. IEEE International Symposium on Personal, Indoor and Mobile Radio Commun. (PIMRC’03), pp. 1162-1165, Sept. 2003.
[12] P. Gao and C. Tepedelenlioglu, “SNR estimation for non-constant modulus constellations,” in Proc. IEEE Wireless Commun. and Networking Conf. (WCNC’04), pp. 24-29, Mar. 2004.
[13] S. M. Key, Fundamenrals of Statisical Signal Processing: Estimation Theory. Englewood Cliffs, NJ: Prentice-Hall, 1993.
[14] R. I. Jennrich and S. M. Robinson, “A Newton-Raphson algorithm for maximum likelihood factor analysis,”Psychometrika, vol. 34, no. 1, pp. 111-123, Mar. 1969.