一种基于SCA的欠定变速跳频信号盲分离算法  

A Underdetermined Blind Source Separation Algorithm of Variable Frequency Hopping Signals Based on Sparse Component Analysis

作　　者：王淼 蔡晓霞 雷迎科 WANG Miao;CAI Xiaoxia;LEI Yingke(Electronic Countermeasure Institute,National University of Defence Technology,Hefei 230037,China)

机构地区：[1]国防科技大学电子对抗学院指挥对抗系,合肥230037

出　　处：《电子信息对抗技术》2020年第5期11-15,共5页Electronic Information Warfare Technology

摘　　要：针对现有欠定变速跳频信号盲分离算法需已知信源数目及算法复杂度高的问题,提出一种基于稀疏分量分析(sparse component analysis,SCA)的“两步法”改进算法。在估计混合矩阵时,结合变速跳频信号时频域稀疏和对高阶累积量不敏感的特点,改进原有聚类算法;将压缩重构匹配追踪思想应用到源信号恢复中。实验证明,改进算法可以得到预期效果,同时在相同条件下,该提算法比传统算法性能优异约6dB。For the present blind signal separation algorithm of the underdetermined variable fre-quency hopping signal needs to know the number of sources and the complexity of the algorithm,a“two-step method”improved algorithm based on sparse component analysis(SCA)is pro-posed.In the estimation of the mixing matrix,combined with the characteristics of the frequency hopping frequency signal sparse frequency and insensitivity to high-order cumulants,the origi-nal clustering algorithm is improved.The compression reconstruction matching tracking idea is applied to source signal recovery.Through experiments,the algorithm is improved to obtain the expected results,under the same SNR.The performance of this algorithm is about 6dB which is better than other algorithms.

关 键 词：欠定盲源分离 变速跳频信号 稀疏分量分析 两步法 

分 类 号：TN911.23[电子电信—通信与信息系统]