基于小波变换的单通道信号盲分离方法  

Single channel signal blind separation method based on wavelet transform

作　　者：解辉 李猛 田建刚 岳夕彪 杨庆培 XIE Hui;LI Meng;TIAN Jiangang;YUE Xibiao;YANG Qingpei(Shijiazhuang Campus of Army Engineering University of PLA,Shijiazhuang 050003,China;Navy Factory 704,Qingdao 266000,China;Unit 32140 of PLA,Shijiazhuang 050003,China;Unit 66389 of PLA,Shijiazhuang 050003,China)

机构地区：[1]陆军工程大学石家庄校区,河北石家庄050003 [2]海军704工厂,山东青岛266000 [3]中国人民解放军32140部队,河北石家庄050003 [4]中国人民解放军66389部队,河北石家庄050003

出　　处：《现代电子技术》2021年第7期56-59,共4页Modern Electronics Technique

基　　金：河北自然科学基金(F2019506037)。

摘　　要：单通道信号的盲分离是非合作信号处理领域的热点问题,文中给出一种利用小波变换有效实现信号分离的方法。首先推导了小波变换的模关于尺度ac具有对称性的结论,进而利用该对称性及小波变换的线性、冗余性和未与其他信号分量重叠的部分尺度的小波变换重构时域对应的信号分量,实现频域部分重合的单通道混合信号的盲分离。仿真结果验证了该算法能够有效实现信号分离,分离效果与信号频谱分离度和输入信噪比成正比,分离结果满足后续对分离信号处理的需求。Blind separation of single-channel signals is a hot topic in the field of non-cooperative signal processing.Under this background,a method of signal separation using wavelet transform is proposed.The conclusion that the wavelet transform modulus is of symmetry on the scale ac is deduced,and then the symmetry,the linearity and redundancy of wavelet transform,and the signal components(corresponding with time domain)reconstructed by partial scale wavelet transform which does not overlap with other signal components are used to realize the blind separation of single-channel mixed signals which is partially-overlapped in frequency domain.The simulation results show that the proposed algorithm can effectively achieve signal separation,and its separation effect is proportional to the signal spectrum separation degree and input SNR.In addition,the separation results can meet the requirements of processing the separated signal subsequently.

关 键 词：信号盲分离 单通道信号 小波变换 通信侦察 信息对抗 信号处理 时频分析 

分 类 号：TN91-34[电子电信—通信与信息系统]