基于压缩感知的信道反馈重构

Channel Feedback Reconstruction Based on Compressed Sensing

在大规模多输入多输出(Multiple-Input Multiple-Output,MIMO)系统中,基于压缩感知技术(Compressed Sensing,CS)开发高效的信道状态信息(Channel State Information,CSI)反馈方案是现在研究的热点。针对现有的基于CS的信道反馈重构算法——正交匹配追踪(Orthogonal Matching Pursuit,OMP)算法存在重构时间长、数据量大可能会无法适用的不足,提出了一种改进的OMP算法,即广义正交匹配追踪(Generalized OMP,GOMP)算法对CSI进行高效重构。仿真结果表明,GOMP算法在重构精确度上高于OMP算法,特别是在较低的压缩比下优势更为突出;而且由于迭代次数减少,需要的重构时间也显著减少。

In massive multiple-input multiple-output(MIMO)systems,the development of efficient channel state information(CSI)feedback scheme based on compressed sensing(CS)is a hot topic.However,the existing CS-based channel feedback reconstruction algorithm,orthogonal matching pursuit(OMP)algorithm,has the disadvantage of taking a great expense of time and even being unavailable in the case of large amounts of data.In this paper,a novel reconstruction algorithm called generalized orthogonal matching pursuit(GOMP)is proposed to reconstruct the compressed CSI efficiently.Simulation results show that the reconstruction accuracy of GOMP is better than that of the OMP,especially at lower compression ratios.Moreover,due to the reduction of iterations,the required reconstruction time is also significantly reduced.