基于循环平稳特性优化的联合算法的盲分离

打开文本图片集

中图分类号：TN911.7 文献标识码：A文章编号：2096-4706（2025）24-0110-06

Abstract：To addressthe isue of low separation accuracy in Blind Source Separation （BSS）algorithms，a combined BSSalgorithmisdevelopedbyintegratingtheJoint Approximate DiagonalizationofEigenmatrices （JADE）and Second-Order Blind Identification（SOBl）algorithms.Furthermore，combined withtheadvantagesof thecyclostationarycharacteristics in thenoisesuppesion，thecycliccumulantsareitroducedintothedesignedjointalgorithm，andajointalgorithmbasedonthe optimizationoftecyclostatioarycharacteristicsisonstructed，whichfurtheroptimizesthealgorithmperformance.ulation resultsdemonstratethattheproposedmethodcaneffectivelyextractthedesiredsignalsandoutperforms traditionalsingle algorithms in terms ofboth separation accuracy and convergence speed，indicating strong potential forapplications.

Keywords：Blind Source Separation; combined algorithm; cyclostationarycharacteristic;cyclic cumulant

0 引言

盲源分离（BlindSourceSeparation）[-2]理论从20世纪80年代提出以来，一直是人们研究的热门课题，其目的是在源信号和传输通道未知的情况下[3-4]，利用天线接收的观测信号提取我们感兴趣的信号。（剩余5597字）