基于压缩感知和Welch的宽带频谱感知新算法  

New Wideband Spectrum Sensing Algorithm Based on Compressed Sensing and Welch

作　　者：王可青[1] 陈建忠[2] 牛英滔[2] 朱勇刚[2] 

机构地区：[1]解放军理工大学通信工程学院研究生3队,江苏南京210007 [2]总参谋部第63研究所,江苏南京210007

出　　处：《军事通信技术》2012年第4期1-5,14,共6页Journal of Military Communications Technology

基　　金：国家自然科学基金资助项目(61102091);中国博士后科学基金资助项目(201003798);江苏省第一批博士后科研计划资助项目(1201076C);解放军理工大学预研基金资助项目(KY63ZLXY1203)

摘　　要：基于压缩感知的谱分析技术是宽带频谱感知的重要方法之一。文章将压缩感知与Welch法相结合,提出了一种新的宽带频谱感知算法。在不重构原信号波形的条件下,直接利用观测矩阵获得的低速观测值进行信号的功率谱估计,降低了采样速率,避免了重构算法的高复杂度。仿真结果表明,与基于压缩感知的多窗口联合奇异值分解算法(CS-MTM-SVD)相比,新算法在压缩率为0.5时能以更高的准确率估计出信号功率谱,估计精度与压缩率的变化成正比关系,并且新算法对噪声容忍度也更高。Spectrum analysis technology based on Compressed Sensing（CS） is one of the im- portant methods for wideband spectrum sensing. Taking both the characteristics in spectrum dis- tribution and the request for spectrum sensing into consideration, a new spectrum sensing algo- rithm was proposed for the wideband cognitive radio based on CS and Welch. It can reduce the sampling rate and avoid the complexity of the reconstruction algorithm under conditions when the original signal is not reconstructed. Simulation results prove the feasibility of the new algorithm, and show that the reconstruction accuracy is proportional to the compression rate and that the new algorithm has higher noise tolerance than the multi-taper method combined with singular- value decomposition based on compressed sensing（CS-MTM-SVD）.

关 键 词：压缩感知 功率谱估计 频谱感知 

分 类 号：TN911.7[电子电信—通信与信息系统]