摘要: |
传统的压缩感知理论考虑的测量值均是实值的,具有无限比特精度。然而在实际应用中,由于数据存储或传输需求,必须考虑测量值的量化问题。通过研究总结已有的一比特量化压缩感知(1-bit Quantized Compressive Sensing)重建算法,提出了一种改进的二进制迭代硬阀值(BIHT)算法。该算法通过引入回溯筛选的思想,在每一步迭代过程中优化了对原子的选择。实验仿真表明,在采样比特数较低时,基于回溯的二进制迭代硬阀值(BBIHT)算法比二进制迭代硬阀值算法重建精度高2~3 dB,且重建速度快。因此,BBIHT算法更有实际应用意义。 |
关键词: 压缩感知 一比特量化 重建算法 |
DOI: |
|
基金项目:武器装备预研基金项目(9140A25031112JB32001);综合业务网理论及关键技术国家重点实验室(西安电子科技大学)开放研究课题(ISN15-13) |
|
An improved 1-bit compressive sensing reconstruction algorithm |
DANG Kui,MA Lin-hua,TIAN Yu,SUN Yu-xue,RU Le |
() |
Abstract: |
The classic compressive sensing(CS) theory assumes the measurements to be real-valued and have infinite bit precision. But in practice,quantization of measurements should be taken into consideration because of the data storage demand or transmission demand. Based on the research and summarization of the existing 1-bit Quantized Compressive Sensing reconstruction algorithms,an improved Binary Iterative Hard Thresholding(BIHT) algorithm is presented. This algorithm optimizes the choice of atoms in each iteration by adding the idea of backtracking. The simulation results show that the reconstruction precision of Backtracking Binary Iterative Hard Thresholding(BBIHT) is 2~3 dB higher than that of BIHT algorithm when the sample bits are not many,besides the reconstruction speed is higher than that of BIHT. So BBIHT is much more efficient and practical. |
Key words: compressive sensing 1-bit quantization reconstruction algorithm |