基于信号分段处理的截尾型序贯检验算法  

A Truncated SPRT Algorithm Based on Signal Segmental Processing

作　　者：吴利平[1] 李赞[1] 李建东[1] 陈晨[1] 

机构地区：[1]西安电子科技大学综合业务网国家重点实验室,陕西西安710071

出　　处：《电子学报》2011年第10期2412-2416,共5页Acta Electronica Sinica

基　　金：国家科技重大专项资金(No.2010ZX03006-002-04);国家自然科学基金(No.61072070);ISN国家重点实验室自主课题(No.ISN1101002);高等学校学科创新引智计划(No.B08038);新世纪优秀人才支持计划(No.NCET-07-0653);长江学者和创新团队发展计划(No.IRT0852)

摘　　要：本文针对复杂电磁环境下认知无线电中的频谱感知需求,结合能量检测算法和序贯似然比检验理论,提出了一种基于信号分段处理的截尾型序贯检验（truncated SPRT）算法.所提算法通过对接收信号进行分段处理和能量计算,并引入中心极限定理,使得序贯检测的似然比函数服从正态分布,从而大大简化后续计算和理论推导过程.为了保证能够在有限检测时间内做出判决并得到最佳检验结果,引入截尾型判决算法,并推导得到了最佳截尾门限.性能分析和蒙特卡罗仿真结果表明：在达到相同检测性能情况下,相比于现有的能量检测算法,所提算法减少了40%~60%的平均容量样本和频谱感知时间,从而保障了复杂电磁环境下未知目标信号的快速、有效感知.According to the spectrum sensing demand under complex electromagnetic environment in cognitive radio（CR） network,a truncated sequential probability ratio test（SPRT） algorithm by utilizing signal proceeding in segment is proposed based on theory and application of energy detection（ED） and SPRT in this paper.By utilizing energy calculation of received signal in segment and introduction of the central limit theorem,the log-likelihood ratio of energy detection approximates Normal distribution,which greatly simplifies the following calculative process and theoretical deduction.Simultaneously,in order to make decision and reach the best test results in a limited sensing time,a truncated SPRT algorithm is introduced and its optimal truncated threshold is derived.Performance analysis and simulation results show that under same performance conditions,the proposed detection reduce average sample number（ASN） and sensing time required around 40%～60%,comparing with the fixed-sample-size（FSS） detection.It guarantees rapid and effective perception of unknown target signal under the complex electromagnetic environment.

关 键 词：截尾型序贯检验 能量检测 频谱感知 认知无线电 复杂电磁环境 

分 类 号：TN911.7[电子电信—通信与信息系统]