基于稀疏傅里叶变换的哈希映射宽带频谱感知算法

发布时间：2018-02-22 14:53

本文关键词： 稀疏傅里叶变换哈希映射频谱分析复杂度宽带频谱感知　出处：《信号处理》2017年08期 　论文类型：期刊论文

【摘要】：宽带频谱感知一般要求对高达数GHz带宽信号进行频谱分析,信号的采样点数大,计算量大。稀疏傅里叶变换算法利用信号频谱稀疏性,高效计算宽带信号频谱,其计算复杂度低于快速傅里叶变换算法。本文详细研究了稀疏傅里叶变换的哈希映射法,证明了频谱重排性质。为了降低频谱漏采的概率,需先对信号进行频谱重排和时域加窗处理;然后进行时域混叠以实现频谱降采样;最后利用哈希反映射和循环投票方法尽可能准确地从降采样的频谱中恢复宽带信号原频谱,从而实现频谱感知。仿真结果表明当采样长度由1024点增加到2048点时,本文方法的运算时间分别比OMP算法减少约19倍和47倍。
[Abstract]:Wideband spectrum sensing generally requires spectrum analysis of up to a number of GHz bandwidth signals, which has a large number of sampling points and a large amount of computation. The sparse Fourier transform algorithm uses the spectral sparsity of signals to efficiently calculate the spectrum of wideband signals. The computational complexity of the algorithm is lower than that of the fast Fourier transform algorithm. In this paper, the hash mapping method of sparse Fourier transform is studied in detail, and the property of spectrum rearrangement is proved. It is necessary to rearrange the signal spectrum and window the signal in time domain, and then the time domain aliasing is carried out to realize the spectrum de-sampling. Finally, the original spectrum of the wideband signal is recovered from the downsampled spectrum as accurately as possible by using the hash reflection and cyclic voting method. The simulation results show that when the sampling length is increased from 1024 to 2048, the computational time of the proposed method is about 19 times and 47 times less than that of the OMP algorithm, respectively.
【作者单位】：杭州电子科技大学通信工程学院;中国电子科技集团第36研究所通信系统信息控制技术国家级重点实验室;
【基金】：“十二五”国防预研项目(41001010401)
【分类号】：TN911.6

【相似文献】