一种自适应压缩感知重构算法  

Adaptive recovery algorithm for compressive sensing

作　　者：甘伟[1] 许录平[1] 罗楠[1] 谢强[1] 

机构地区：[1]西安电子科技大学电子工程学院,陕西西安710071

出　　处：《系统工程与电子技术》2011年第9期1948-1953,共6页Systems Engineering and Electronics

基　　金：国家自然科学基金(60772139);国家高技术研究发展计划(863计划)(2007AA12Z323);教育部高等学校博士学科点专项科研基金(200807011007)资助课题

摘　　要：为优化稀疏自适应匹配追踪(sparsity adaptive matching pursuit,SAMP)算法的性能,给出了一种修正自适应匹配追踪(modified adaptive matching pursuit,MAMP)算法。该算法采用模糊阈值预选方案,改进了步长选择方法,设置了初次裁剪门限。仿真结果表明,在同等稀疏的条件下实现精确重构,该算法的运算速度较原算法提高了1倍,所需的观测值个数减少了1%,并提高了重构精度。In order to optimize the performance of sparsity adaptive matching pursuit （SAMF） algorithm, a modified adaptive matching pursuit （MAMP） algorithm is proposed. The proposed algorithm adoptes a fuzzy threshold preliminary rule, improves the method of choosing step size and sets the initial threshold to cut out the candidate set. Simulation results show that, compared with SAMP, the operation speed of MAMP is increased by a factor of 2 for the same sparsity level, and the required measurement number decreases about 1 %. In addition, the recovery accuracy is also improved.

关 键 词：压缩感知 信号重构 自适应匹配追踪 模糊阈值 

分 类 号：TN911.72[电子电信—通信与信息系统]