混合最小化问题下重构块稀疏信号的充分条件  

A Sufficient Condition on Recovery of Block Sparse Signals via Mixed Minimization

作　　者：卜京 王金平 BU Jing;WANG Jinping(School of Mathematics and Statistics,Ningbo University,Ningbo 315211,Zhejiang,China)

机构地区：[1]宁波大学数学与统计学院,浙江宁波315211

出　　处：《武汉大学学报（理学版）》2021年第2期185-189,共5页Journal of Wuhan University:Natural Science Edition

基　　金：国家自然科学基金(62071262)。

摘　　要：基于混合最小化问题,提出了测量矩阵的块鲁棒零空间特性(robust null space property,rNSP),在有噪声情况下证明rNSP是稳定重构块稀疏信号的充分条件。并且当块限制等距常数小于0.453 1时,证明块限制等距性蕴含着块鲁棒零空间特性。This paper focuses on the mixed minimization case and proposes the block robust null space property(rNSP)of measurement matrix, and then proves that the block rNSP is a sufficient condition to stably recover block sparse signals in noisy systems. Then we prove that the block restricted isometry property implies the block rNSP when block restricted isometry constant is less than 0.453 1.

关 键 词：块鲁棒零空间特性 块限制等距性 混合最小化 块稀疏 

分 类 号：O177.92[理学—数学]