正交分頻多工通訊系統未知與非時變衰減通道之識別與估測

Abstract

本篇論文為討論正交分頻多工(OFDM)通訊系統在使用最大概似通道估測(MLE)法於不同的非時變通道環境如連續型，叢集型，兩根多重路徑的通道型態下的性能表現，並利用估測結果作進一步的通道等化以改進系統效能。 根據我們所掌握的部分通道資訊我們可以藉由引進通道功率估測臨界值(channel power detection threshold)作為多重路徑估測判別依據而增加最大概似通道估測的準確度，並在模擬的結果中觀察到可以獲得2~3 dB的通訊系統性能改善。 最後，在通道資訊完全未知的情況下，藉由傳送訓練序列中的前置符碼(Preamble)架構，我們提出平均視窗尺度演算法(Averaged window size algorithm)來幫助我們決定出最大概似通道估測所需的通道統計特性資訊，並以錯誤偵測率為評量指標探討其性能表現。

This paper is talk about the performance evaluation under different deterministic time-invariant channel profiles like as continuous, cluster, two path channel profiles using maximum likelihood channel estimation (MLE) scheme in Orthogonal Frequency Division Multiplexing (OFDM) communication system to obtain the channel information and utilizing it for further equalization. According to the Channel statistics we have to improve the maximum likelihood channel estimation accuracy under partial known channel information by involving a channel power detection threshold to help us deal with path estimation and gain a 2~3 dB system performance improvement. We can find the result in corresponding simulations. Finally, we propose an Averaged Window size algorithm to decide the channel statistics needed for MLE channel estimation by transmitting preamble in training sequence under unknown channel profile and discuss the performance of the Average Window size algorithm with error detection probability as an evaluation index.