用于随机等效采样的多频段信号重建

Multiband signal reconstruction for random equivalent sampling.

作者信息

Zhao Y J, Liu C J

机构信息

School of Automation Engineering, University of Electronic Science and Technology of China, Chengdu 611731, China.

出版信息

Rev Sci Instrum. 2014 Oct;85(10):105109. doi: 10.1063/1.4899204.

DOI:10.1063/1.4899204

PMID:25362458

Abstract

The random equivalent sampling (RES) is a sampling approach that can be applied to capture high speed repetitive signals with a sampling rate that is much lower than the Nyquist rate. However, the uneven random distribution of the time interval between the excitation pulse and the signal degrades the signal reconstruction performance. For sparse multiband signal sampling, the compressed sensing (CS) based signal reconstruction algorithm can tease out the band supports with overwhelming probability and reduce the impact of uneven random distribution in RES. In this paper, the mathematical model of RES behavior is constructed in the frequency domain. Based on the constructed mathematical model, the band supports of signal can be determined. Experimental results demonstrate that, for a signal with unknown sparse multiband, the proposed CS-based signal reconstruction algorithm is feasible, and the CS reconstruction algorithm outperforms the traditional RES signal reconstruction method.

摘要

随机等效采样（RES）是一种采样方法，可用于以远低于奈奎斯特速率的采样率捕获高速重复信号。然而，激励脉冲与信号之间时间间隔的不均匀随机分布会降低信号重建性能。对于稀疏多频带信号采样，基于压缩感知（CS）的信号重建算法能够以极高的概率梳理出频带支撑，并减少RES中不均匀随机分布的影响。本文在频域中构建了RES行为的数学模型。基于所构建的数学模型，可以确定信号的频带支撑。实验结果表明，对于具有未知稀疏多频带的信号，所提出的基于CS的信号重建算法是可行的，并且CS重建算法优于传统的RES信号重建方法。