利用组稀疏特性的宽带压缩频谱感知

发布时间：2018-02-09 07:22

本文关键词： 宽带频谱感知压缩感知组稀疏贪婪算法　出处：《信号处理》2014年03期 　论文类型：期刊论文

【摘要】：基于认知无线电的动态频谱接入需对宽带信道进行频谱感知,而越来越高的采样速率日益成为宽带频谱感知的瓶颈。压缩感知作为一种新的信号获取技术为亚奈奎斯特采样速率下的宽带频谱感知提供了一种可行方案。在相关应用场景中,如果能够挖掘相关先验信息并在重构算法中整合这些信息,将大幅提高压缩感知的性能。本文基于压缩感知技术,利用信道的划分信息及宽带信号的组稀疏特性,提出了一种组稀疏贪婪算法GOMP。该方法在成熟的贪婪算法基础上,利用子信道内多频点的组测量信息,根据组测量的概率分布特性来识别宽带信道的活动子信道。这种组测量识别方式使算法能以较少的观测数据实现对宽带信道的快速准确感知,极大地降低了宽带频谱感知所需的采样速率。实验结果表明:该算法比传统的OMP算法及BP算法不仅具有更好的重构效果及频谱检测性能,而且具有更好的压缩性能及实时性能。
[Abstract]:Dynamic spectrum access based on cognitive radio requires spectrum sensing in broadband channels. As a new signal acquisition technology, compressed sensing provides a feasible scheme for broadband spectrum sensing at Nyquist sampling rate. If we can mine the relevant prior information and integrate the information in the reconstruction algorithm, we will greatly improve the performance of compression sensing. Based on the compression sensing technology, this paper makes use of the channel partition information and the sparse characteristics of broadband signals. In this paper, a group sparse greedy algorithm, GOMP. based on the mature greedy algorithm, is proposed. The active subchannels of broadband channels are identified according to the probability distribution characteristics of group measurements, which enable the algorithm to realize fast and accurate perception of wideband channels with less observation data. The experimental results show that the proposed algorithm not only has better reconstruction effect and spectrum detection performance than the traditional OMP algorithm and BP algorithm, but also has better compression performance and real-time performance.
【作者单位】：华中科技大学电子与信息工程系;长沙理工大学计算机与通信工程学院;
【基金】：国家自然科学基金(60933012)
【分类号】：TN925

【参考文献】