摘要
由于多载波系统无线信道固有的稀疏特性,压缩感知技术(compressed sensing,CS)已被应用于正交频分复用(orthogonal frequency division multiplexing,OFDM)系统的信道估计中以提高频谱利用率。然而,传统的时域普通采样方法会导致信道恢复字典不够精细,无法精确反映传输信道路径特性。针对这一问题,提出采用多径稀疏分数时延信道模型来模拟OFDM系统的无线多径信道,利用在接收端进行时域过采样方法细化信道恢复字典以提高信道估计精度。同时,针对过采样引起的压缩感知测量矩阵的扩大而导致重构算法的复杂度增加的问题提出采用广义正交匹配追踪算法(generalized orthogonal matching pursuit,GOMP)以降低计算复杂度。仿真结果表明接收端时域过采样方法能准确检测到信道的分数时延且采用的GOMP算法能将传统的OMP算法的复杂度降低近80%,验证了所采用的信道估计方法的可靠性和有效性。
Due to the inherent sparse feature of the wireless channel in multi carrier system,compressed sensing techniques are applied in the channel estimation of the OFDM system to improve spectrum utilization.However,the conventional time domain sampling method causes the imprecision to the channel recovery dictionary,which can’t accurately reflect the characteristics of the transmission channel path.To solve this problem,a multipath sparse fractional delay channel model is used to simulate the OFDM wireless multipath channel.In addition,a time domain oversampling based generalized orthogonal matching pursuit(GOMP)algorithm at the receiver is exploited to improve the estimation accuracy and keep a relatively low computational complexity which is increased due to the time domain oversampling.Simulation results demonstrate that the time domain oversampling can effectively detect the fractional delay.In contrast to the conventional OMP algorithm,the GOMP algorithm reduces the computational complexity by80%.Hence,it verifies the reliability and the validity of the proposed method in this paper.
作者
马子骥
彭强
周冰航
李元良
唐涛
MA Ziji;PENG Qiang;ZHOU Binghang;LI Yuanliang;TANG Tao(School of Electrical and Information Engineering, Hunan University, Changsha 410082, P.R. China)
出处
《重庆邮电大学学报（自然科学版）》
CSCD
北大核心
2017年第5期611-617,共7页
Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)
基金
中央国有资本经营预算项目([2013]470号)
中央高校基本科研项目(2014-004)
国家自然科学基金(61540012)
中国博士后科研基金(2014M562100)
湖南省科技计划重点项目(2015JC3053)
教育部产学合作协同育人项目(201601004010)~~
关键词
压缩感知
匹配追踪
分数时延
过采样
compressed sensing
matching pursuit
fractional delay
oversampling
