非参数似然率独立分量分析算法的OFDM系统载波频偏盲估计  

Blind estimation of OFDM system carrier frequency offset using independent component analysis algorithm with nonparametric likelihood ratio

作　　者：陈朝阳[1] 邢海涛[1] 杨光松[1] 马中华[1] 林子杰[1] 

机构地区：[1]集美大学信息工程学院,厦门361021

出　　处：《仪器仪表学报》2011年第9期1967-1972,共6页Chinese Journal of Scientific Instrument

基　　金：福建省教育厅科技项目(No.JA07137);福建省厦门市科技局项目(No.3502Z20093021)资助

摘　　要：研究提出了一种利用非参数似然比(NLR)新算法获得OFDM系统频偏估计的新方法,给出基于NLR的ICA算法的OFDM系统数学模型,利用NLR算法分离各个子载波,然后利用本地载波与子载波的频差估计出OFDM的频偏(CFO),该算法获得的CFO可以包括信道的传输特性对CFO的影响。经典的ICA算法(如FAST-ICA)对源信号的统计特性具有依赖性,源信号统计特性的变化可能使算法的性能降低甚至无法得到希望的分离信号,非参数似然算法(NLR)的独立分量分析(ICA),不依赖于源信号的统计特性,而且能够对混合信号实行连续分离,因此是一个全盲的算法。仿真结果表明,将跟踪到的频偏在接收端进行补偿后,减小了子载波的串扰,降低系统解调后的误码率,提高了OFDM系统性能。In this paper we present a new method for estimating OFDM system carrier frequency offset （CFO） using a nonparametric likelihood ratio （NLR） algorithm. We bring forward an OFDM mathematic model for ICA separation algorithm based on NLR. The method separates all subcarriers then estimates the OFDM carrier frequency offset through calculating the frequency difference between the separated subcarriers and local carriers. The CFO estimation results calculated using the new method include channel characteristic affection. Classical ICA algorithms （such as FAST- ICA） rely on the assumptions of the source signal statistics; these algorithms may perform sub-optimally or even fail to produce the desired source separation when the assumed source statistics model is inaccurate. However the proposed NLR algorithm is completely blind to sources and has the ability to separate the mixed sources continuously. The proposed algorithm does not rely on source statistics. From the simulation results, we conclude that after compensating CFO on the receiver, ICI diminishes, BER decreases and the performance of the OFDM svstem is improved.

关 键 词：载波频率偏移 非参数似然比 独立分量分析 正交频分复用 

分 类 号：TN911.7[电子电信—通信与信息系统]