標題: 離散多重音調接收機之低複雜度等化演算法
Low-complexity Equalization Algorithms for DMT Transceiver
作者: 張杰碩
Chieh-Shuo Chang
吳文榕
Wen-Rong Wu
電信工程研究所
關鍵字: 離散多重音調系統;DMT
公開日期: 1998
摘要: 目前現有之離散多重音調系統之等化器參數求取演算法具接近最佳的表現,但是其計算量相當複雜。以頻域適應性演算法求取參數有收斂不穩定的情形,而時域適應性演算法的收斂速度有待提升。本論文即針對這些問題提出解決的方法。首先,對於時域的適應性演算法我們提出一種能加速運算時間的架構。主要是在接收端加入一前置濾波器以縮短通道的長度,而前置濾波器的係數則由已知通道利用最佳化的方法求出。接著我們提出一種新的演算法能快速計算等化器參數的理論值。我們實地以計算機模擬了兩者的表現並與原來的演算法比較。結果發現我們所提的方法之效能和傳統的做法很接近,但運算量卻能大幅的減低。
Existing algorithms for the computation of DMT equalizer coefficients require, though have near-optimum performance, high computation load. Frequency-domain equalizer training lacks the stability of convergence, and time-domain equalizer training suffers low speed of convergence. We proposed in this paper two alternative algorithms, one of which is to speed up the convergence of time-domain euqalizaer training, the other is to reduced the computation load of coefficients calculation. The performance is simulated and compared to that of the original algorithms.
URI: http://140.113.39.130/cdrfb3/record/nctu/#NT870435036
http://hdl.handle.net/11536/64495
Appears in Collections:Thesis