正交分頻多工系統對高速移動通道中的子通道間干擾抑制

Abstract

近年來，正交分頻多工在高速傳輸上的應用變的愈來愈普及。由於正交分頻多工相對於單載波系統具有相當長的符號區間，故能夠有效地對抗幅間干擾，在頻率選擇性通道中相當的強健。然而對於高速移動的應用下，符號區間內通道的變化將導致接收端的子通道間干擾而破壞子通道間的正交性。而該子通道間干擾隨著移動速度、載波頻率、和符號區間長度增加而增加。此干擾導致效能的衰減與錯誤地板(error floor)的產生。 文獻的方法，基於平行干擾消除與通道脈衝響應的片段線性化內插，已被提出。然而在高都普勒延展的情況下，片段線性化的假設逐漸的偏離。在這裡，提出三種通道脈衝響應的重建方法來取代片段線性化，包含：二階內插、低通濾波器內插、和溫氏濾波器內插。並模擬四種方法下的效能比較。從模擬結果得知，二階內插在高訊雜比的狀況下會比片段線性化的效能好，而低通濾波器由於能夠濾除雜訊有更好的效能。溫氏濾波器由於利用了通道的關聯矩陣，在四種演算法中有最佳的效能。

In the recent years, OFDM has become one of the most popular transmission technologies for broadband wireless communication systems because of the ability of combating Inter-Symbol Interference(ISI) in the frequency selective channel and the excellent bandwidth efficiency. However, when OFDM systems are applied in the high speed mobile applications, the channel time-variation, the namely time selectivity in a symbol duration introduces Inter-carrier interference (ICI), which destroys the orthogonal property between sub-carriers. ICI will become severe when the mobile speed, the carrier frequency, and the symbol duration increase. Without solving this problem, the receiving performance will be degraded, and the bit error floor will appear even in the high SNR condition. The previously published method of solving the mobile ICI problem is composed of the parallel interference cancellation (PIC) and the time-varying channel frequency response reconstruction (estimation). The reconstruction is based on piece-wise linear interpolation in time. But, in the high mobility wireless receiving, the piece-wise linearity assumption of channel response will be not true. Using the piece-wise linear interpolation for channel estimation cannot obtain satisfactory performance in the high mobility condition. In this thesis, we propose three interpolation algorithms for mobile channel reconstruction. They are the second order interpolation, the low-pass filter interpolation, and the Wiener filter interpolation. The simulation results show that the proposed methods are better than the previously published method especially at high SNR. Utilizing the information of channel correlation matrix, the Wiener filter can provide the best performance.