標題: 多輸入多輸出正交分頻多工系統中Clipping Noise之消除
Clipping Noise Mitigation in MIMO-OFDM Systems
作者: 俞丁發
Ding-Fa Yu
吳文榕
Dr. Wen-Rong Wu
電信工程研究所
關鍵字: 多輸入多輸出正交分頻多工;切割雜訊;峰均值比;選擇性對應;Clipping Noise;MIMO-OFDM;V-BLAST;PAPR;SLM
公開日期: 2004
摘要: 在正交分頻多工(OFDM)系統中,過高的時域訊號峰均值比(PAPR)會造成傳送訊號直流偏壓過高,使其增加傳送功率之消耗。採以限制訊號振幅的方式來達到降低訊號峰值是既簡單且有效率的解決方式。然而,這會引進額外的雜訊(Clipping Noise, CN),進而嚴重影響系統的表現。雖然消除CN的方法在單天線系統已被廣泛的討論與研究,但是在多天線系統下此問題依舊存在且甚少有相關的研究。本論文即著重於解決此問題。藉由使用V-BLAST的軟性與硬性決策輸出,我們可以延伸傳統消除CN的方法到MIMO-OFDM系統上,如DAR與CNEC。另外,對於DAR與CNEC演算法而言,CR(Clipping Ratio)必須是已知的。在實際應用上,接收端可能無法獲得此資訊。於是我們提出在MIMO-OFDM系統中,一個簡單且有效估測CR的方法。為了更進ㄧ步降低訊號峰均值比,我們使用傳送端選擇性對應(Selective Mapping)的架構。最後,我們以無限區域網路IEEE 802.11n為模擬平台,藉由模擬結果發現,我們提出的方法可以有效降低MIMO-OFDM系統下的訊號峰均值比,且估計CR效能也相當令人滿意。
OFDM signal exhibits a high peak to average power (PAPR) ratio, which leads to a high amplifier power consumption. Clipping is known to be a simple and effective method to solve the problem. However, clipping will introduce clipping noise reducing the system performance. Although clipping noise mitigation for single-input and single-output (SISO) OFDM systems has been studied, that for MIMO-OFDM system is rarely considered. This thesis is aimed to solve the problem. Using soft and hard decisions from the V-BLAST algorithm, we are able to extend the conventional clipping noise mitigation methods such as the DAR and the CNEC scheme to MIMO-OFDM systems. One requirement for DAR and CNEC algorithm is that the clipping ratio (CR) has to be known. This may not always be possible in practice. We then propose a simple yet effective CR estimation method for MIMO-OFDM systems. To further reduce the PAPR, we also apply the transmitter selective mapping (SLM) scheme. Finally, we use an IEEE802.11n WLAN system as our simulation platform. The results show that the proposed method can significantly reduce the PAPR of MIMO-OFDM systems and our CR estimation method performs satisfactorily.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT009213514
http://hdl.handle.net/11536/69557
Appears in Collections:Thesis


Files in This Item:

  1. 351401.pdf