一种改进的基于平滑L_(0)范数的压缩感知重构算法  

An Improved Compressed Sensing Reconstruction Algorithm Based on Smooth L_(0)Norm

作　　者：徐刚 陈斌辉[1] 付卫红[2] XU Gang;CHEN Bin-hui;FU Wei-hong(No.36 Research Institute of CETC,Jiaxing,Zhejiang 314033,China;School of Telecommunication Engineering,Xidian University,Xi'an,Shannxi 710071,China)

机构地区：[1]中国电子科技集团公司第三十六研究所,浙江嘉兴314033 [2]西安电子科技大学通信工程学院,陕西西安710071

出　　处：《通信对抗》2024年第4期17-21,共5页Communication Countermeasures

摘　　要：针对现有的基于平滑L_(0)范数的重构算法重构精度低、重构速度慢的问题,提出了一种改进的基于平滑L_(0)范数的压缩感知信号重构算法。该算法首先利用基于修正牛顿法的SL_(0)(NSL_(0))算法求得一组近似解,再利用贪婪算法的思想得到精确解。仿真实验表明,当压缩比为0.6,信噪比为15dB时,所提算法重构信号的信干比比NSL_(0)算法的提高了6dB,算法运算时间减少了40%,因此所提算法在保证重构精度的同时,降低了计算复杂度,提高了运算速度。Aiming at the problem that the existing reconstruction algorithm based on smooth L_(0) norm has low accuracy and slow reconstructing speed,an improved SL_(0) reconstruction algorithm in compressed sensing was proposed in this paper.The algorithm firstly get a set of solutions using the modified Newton smooth L_(0) norm(NSL_(0))algorithm,then the idea of greedy algorithm is used to get the exact solutions.The Simulation results show that the signal-to-interference ratio of reconstructed signal obtained by the proposed algorithm is 6dB higher than that of NSL_(0) algorithm and the algorithm operation time is reduced by 40%in the condition that the SNR is 15dB and compression ratio is 0.6.Therefore,the proposed algorithm not only ensures the reconstruction accuracy,but also reduces the computational complexity of the algorithm and improves the algorithm's operation speed.

关 键 词：压缩感知 重构算法 平滑L_(0)范数 修正牛顿法 贪婪算法 

分 类 号：TN911.7[电子电信—通信与信息系统]