在電力線傳輸環境裡，充滿各式各樣的脈衝雜訊，這些脈衝雜訊與傳統的AWGN雜訊不同點在於，脈衝雜訊的能量往往是AWGN的數百倍，而常見的脈衝雜訊有Class A model和Bernoulli-Gaussian通道模型，而這兩種雜訊皆屬於無記憶性型，發生雜訊的時機非常隨機，無法去描述真實通道的特性，故發展出基於馬可夫鏈特性的雜訊Markov-Gaussian(MG)通道模型。
在本文中，使用維特比演算法去做脈衝雜訊的偵測，由於馬可夫高斯雜訊有狀態間的轉移機率，利用這點特性與格狀圖相似，故能使用維特比演算法進行，並使用上一個時間點的資訊去計算通道狀態。另外本所用的裁剪是為了抵抗能量過大的脈衝雜訊，讓解碼端易於資料更正。在模擬結果中顯示運用維特比演算法在訊雜比7.5dB時就能達到平均位元錯誤率10-5之區間，而我們所推導的在只知道部分資訊下的偵測器，與最佳偵測器的結果是非常相近的

There are many kinds of impulsive noise in the power line communication and they always have stronger energy than Additive White Gaussian Noise. There are common impulse noise such as Middleton class A and Bernoulli-Gaussian noise model. Both of them are memoryless, which mean that their occurrences are random. However they can’t describe the characteristics of the real channel. A memory channel such as Markov-Gaussian is introduced to address the characteristics of the real channel.
In this thesis, a self-arbitrating mechanism is introduced to detect the presence of impulse noise, which is in the bad state. Because transition state of Markov-Gaussian channel is similar with trellis, therefore the Viterbi Algorithm could be used to compute the previous time instant metric and decide the channel state. A clipping method is used to help the decoder to recover the information. Compared to the other methods, the proposed algorithm has 7.5 dB gains in terms of Signal to Noise Ratio (SNR) at the Bit Error Ratio (BER) value of 10-5. The simulation result indicated that the proposed Viterbi decoding schemes is robust: the BER attained using the proposed decoders is remarkably close to that of an optimal and refined decoder which uses impulse statistics.

[1]M. Zimmermann and K. Dostert, "Analysis and modeling of impulsive noise in broad-band powerline communications," IEEE Transactions on Electromagnetic Compatibility vol. 44, pp. 249-258, 2002.
[2]K. L. Blackard, T. S. Rappaport, and C. W. Bostian, "Measurements and models of radio frequency impulsive noise for indoor wireless communications," IEEE Journal on Selected Areas in Communications, vol. 11, pp. 991-1001, 1993.
[3]P. Banelli, "Bayesian Estimation of a Gaussian Source in Middleton's Class-A Impulsive Noise," IEEE Signal Processing Letters, vol. 20, pp. 956-959, 2013.
[4]M. Ghosh, "Analysis of the effect of impulse noise on multicarrier and single carrier QAM systems," IEEE Transactions on Communications, vol. 44, pp. 145-147, 1996.
[5]M. Mushkin and I. Bar-David, "Capacity and coding for the gilbert-elliott channels," IEEE Transactions on Information Theory, vol. 35, pp. 1277-1290, 1989.
[6]S. Galli, "A Novel Approach to the Statistical Modeling of Wireline Channels," IEEE Transactions on Communications vol. 59, pp. 1332-1345, 2011.
[7]C. Berrou, A. Glavieux, and P. Thitimajshima, "Near Shannon limit error-correcting coding and decoding: Turbo-codes. 1," in Communications, 1993. ICC '93 Geneva. Technical Program, Conference Record, IEEE International Conference on, 1993, pp. 1064-1070 vol.2.
[8]T. Der-Feng, Y. S. Han, M. Wai Ho, C. Li-Chung, and A. J. H. Vinck, "Robust Clipping for OFDM Transmissions over Memoryless Impulsive Noise Channels," IEEE Communications Letters, vol. 16, pp. 1110-1113, 2012.
[9]T. Der-Feng, F. G. Mengistu, Y. S. Han, M. Abera Mulatu, C. Li-Chung, and T. Tzung-Ru, "Robust Turbo Decoding in a Markov Gaussian Channel," IEEE Wireless Communications Letters, vol. 3, pp. 633-636, 2014
[10]D. Fertonani and G. Colavolpe, "On reliable communications over channels impaired by bursty impulse noise," IEEE Transactions on Communications, vol. 57, pp. 2024-2030, 2009.