基于序贯压缩感知的自适应宽带频谱检测  

Adaptive wide-band spectrum detection based on sequential compressed sensing

作　　者：顾彬[1,2] 杨震[1,2] 胡海峰[1,2] 

机构地区：[1]南京邮电大学信号处理与传输研究院,南京210003 [2]南京邮电大学"宽带无线通信与传感网技术"教育部重点实验室,南京210003

出　　处：《仪器仪表学报》2011年第6期1272-1277,共6页Chinese Journal of Scientific Instrument

基　　金：国家高技术研究发展计划(2009AA01Z241);国家自然科学基金(60971129;61071192)资助项目

摘　　要：信号信息处理的新技术压缩感知,提供了认知无线电中宽带频谱检测的新方案。该方案的一个主要问题是在不知道信号稀疏度的前提下如何确定所需要的随机观测次数。提出一种基于序贯压缩感知的自适应检测方案。次用户利用模拟/信息转换器输出观测序列并进行自相关运算,相邻两次自相关观测向量输入融合中心经由SSAMP协作重构算法估算重构误差,并自适应确定所需的最小观测次数。精确重构的稀疏向量利用基于置信度的DS融合算法对主用户宽带频谱占用做出决策。仿真结果证实了误差估计及自适应协作检测的有效性。同时,SSAMP算法与经典的SOMP算法在重构均方误差和检测时间两方面进行了对比。Compressed sensing is a novel technology in signal information processing.It offers a new wide-band spectrum detection scheme in cognitive radio.A major challenge of this scheme is how to determinate the required number of measurements while the signal sparsity is not known a priori.This paper presents an adaptive detection scheme based on sequential compressed sensing in which sequential measurements are collected from the analog-to-information converters of secondary users.The autocorrelation vector of two adjacent measurements is sent to fusion center where a novel cooperative compressed sensing recovery algorithm named SSAMP is utilized for sequential compressed sensing in order to estimate the reconstruction errors and adaptively determinate the required minimal number of measurements.The reconstruction sparse vectors are then used to make a decision about spectrum occupancy of primary users based on credibility and DS theory.Simulation results corroborate the effectiveness of error estimation and adaptive detection performance of our cooperative scheme.Meanwhile,the performances of MSE and detection time for SSAMP algorithm and classical SOMP algorithm are compared and evaluated.

关 键 词：认知无线电 频谱检测 压缩感知 匹配追踪 

分 类 号：TN92[电子电信—通信与信息系统]