基于簇的块稀疏压缩感知的60GHz信道估计.pdfVIP

信怠传输」每接入拉术 基于簇的块稀疏压缩感知的60 GHz 信道估计 侯猛，李斌，孙学斌，赵成林，周正 (北京邮电大学无线网络实验室，北京 100876) 摘 要:基于压缩感知设计适用于 60 GHz 毫米波通信系统的信道估计方案，深入研究了正交匹配追踪( OMP) 算 法和正则正交匹配追踪( Regularized OMP) 算法的 60 GHz 信道估计性能;在此基础上，充分发掘 60 GHz 无线多径信道 所呈现出的分簇特性，提出一种新颖的基于簇分级的稀疏压缩感知重构算法。新算法在有效减少重构迭代次数的前 提下，亦能显著降低信道估计误差。综合对比分析了基于簇分块稀疏压缩感知重构算法和现有压缩感知算法在 60GHz 1言道估计应用中的重构性能，仿真结果表明，压缩感知算法可有效应用于 60 GHz 系统信道估计，而新设计的基 于簇分级的稀疏压缩感知算法则在估计精度和实现复杂度方面具更优越性能。 关键词: 60 GHz; 压缩感知;信道估计 ;CS-ROMP; 分级 中图分类号: TN928 文献标识码:A 文章编号: 1003 一 3114(2012)俩一32 - 3 60 GHz Channel Estimation Based on Cluster-classification and Block-sparse Compressed Sensing HOU Meng ,LI Bin , SUN Xue-bin ,ZHAO Cheng-lin ,ZHOU Zheng (Wireless Network Lab ,ßeijing University of Posts and Telecommunications ,Beijing 100876 ,China) Abstract:The application of CS-OMP and CS-ROMP in channel estimation of 60 GHz millimeter-wave communication systems is investigated. Then the clustering characteristics of 60 GHz wireless multipath channel are fully exploited , based on which , a cluster- classification and block-sparse compressed sensing algorithm is proposed. The new algorithm significantly reduces the reconstruction error of channel estimation under the premise of less iteration times. Error ratios of CS-OMP , CS-ROMP and the proposed algorithm are compared through simulation. The results indicate that CS-OMP and CS喃 ROMP algorithm can be used in 60 GHz channel estimation effectively. However ,the cluster-classification and block-sparse compressed sensing algorithm has a superior performance in channel estimation. Key words :60 GHz; compressed sensing; channel estÌmation; CS-ROMP; cluster-classification 号重构[110 。 引言 标准的 CS 理论框架并未考虑源信号的结构，