摘要
压缩感知理论(Compressed Sensing,CS)为宽带信号的直接信息采样(Analog-to-Information,A-to-I)提供了解决方法。针对现有的CS压缩算法没有充分考虑信息的重要性差异,运用仿真手段验证了多分辨率压缩感知(Multiresolution CS,MCS)思想的正确性,仿真结果表明:MCS能够动态地保护信号中蕴含的重要信息,且与单一分辨率CS重构算法比较,精度明显提高。为下一步将MCS为直接信息采样提供软件支持。
The compressed sensing（CS） theory provides a golden opportunity for the analog-to-information（A-to-I） sampling of wideband signal.However,the existing CS algorithms neglect the diversity in the information significance.This paper validates the fundamental idea of multiresolution compressed sensing（MCS） by simulation method.The simulation results show that MCS can protect key information dynamically and improve significantly the accuracy compared to the single resolution reconstruction algorithm,and will pave the way for A-to-I convertor realization in the future.
出处
《无线电通信技术》
2011年第3期41-43,共3页
Radio Communications Technology
基金
"泰山学者"建设工程专项经费资助
关键词
压缩感知
多分辨率
多尺度
直接信息采样
compressed sensing
multiresolution
multiscale
analog-to-information
