摘要
正交匹配追踪(Orthogonal Matching Pursuit,OMP)算法是压缩感知系统中应用最广泛的重构算法之一.OMP算法中的最小二乘(Least Squares,LS)问题涉及到矩阵求逆运算,是计算复杂度最高的部分.本文提出了一种近似OMP算法,通过优化最小二乘问题提高重构速度,更利于硬件实现.设计了一种基于FPGA的近似OMP算法的硬件实现架构,主要包含计算模块、存储模块和控制模块3个部分,利用Xilinx公司的Vivado软件进行综合仿真实验.对比现有相似实现设计,本文设计的架构通过调用更多的DSP资源减少了slices资源的使用量,在208MHz频率下重构速度提升了1.25-1.72倍,同时具备一定的可扩展性.
Orthogonal Matching Pursuit(OMP)algorithm is one of the most widely used reconstruction algorithms in compressed sensing systems.The least squares(LS)problem in OMP algorithm involves matrix inversion operation,which is the most computationally complex part.In this paper,an approximate OMP algorithm is proposed to improve the reconstruction speed by optimizing the least square problem,which is more conducive to hardware implementation.A hardware architecture of approximate OMP algorithm based on FPGA is proposed,which mainly includes calculation module,memory module and control module.The simulation experiment is carried out by using Vivado software of Xilinx.Compared with the existing similar implementation design,the proposed structure reduces the usage of slices by calling more DSP resources.The proposed structure is reconfigurable and 1.25-1.72 faster in reconstruction time at 208 MHz.
作者
张金鹏
钱慧
王仁平
ZHANG Jin-peng;QIAN Hui;WANG Ren-ping(School of Physics and Information Engineering,Fuzhou University,Fuzhou 350116,China)
出处
《小型微型计算机系统》
CSCD
北大核心
2022年第4期707-711,共5页
Journal of Chinese Computer Systems
基金
数字福建物联网工程应用实验室建设项目(82917002)资助。
关键词
压缩感知
正交匹配追踪算法
最小二乘
最大值计算
FPGA
compressed sensing
orthogonal matching pursuit algorithm
least squares
maximum calculation
FPGA