中国空间科学技术 ›› 2013, Vol. 33 ›› Issue (1): 61-68.doi: 10.3780/j.issn.1000-758X.2013.01.010

• 技术交流 • 上一篇    下一篇

基于SSA和BSS的单通道盲信号分离算法

于宁宇1,马红光1,姜勤波1,石磊2   

  1. (1 第二炮兵工程大学,西安 710025)(2 解放军96656部队,北京 100028)
  • 收稿日期:2011-11-02 修回日期:2012-07-10 出版日期:2013-02-25 发布日期:2013-02-25
  • 作者简介:于宁宇 1983年生,2012年获第二炮兵工程大学火炮、自动武器与弹药工程专业博士学位。研究方向为电子侦察、单通道盲信号分离技术。
  • 基金资助:

    国家自然科学基金(61174207,61074072)资助项目

SSA and BSS Based Separation Algorithm for Single-channel Signals

Yu Ningyu1,Ma Hongguang1,Jiang Qinbo1,Shi Lei2   

  1. (1 The Second Artillery Engineer College, Xi'an 710025) (2 Unit 96656 of PLA, Beijing 100028)
  • Received:2011-11-02 Revised:2012-07-10 Published:2013-02-25 Online:2013-02-25

摘要: 针对空间侦察中单通道宽带接收机截获到多个独立辐射源信号时的盲源分离问题,提出了一种新的单通道盲信号分离算法。该算法首先利用奇异谱分析(SSA)构建伪阵列信号,进而采用盲源分离算法(BSS)实现信号分离。仿真试验表明:该算法可以有效地分离空间侦察中的几种常用信号,对单频信号和线性调频信号构成的单通道信号,在信噪比大于6dB时,分离前后信号的相似系数大于0.9;对不同的相移键控信号构成的单通道信号,在信噪比大于4dB时,误码率小于0.01。

关键词: 信号处理, 奇异谱分析, 单通道盲信号分离, 空间电子侦察

Abstract: In blind source separation(BSS),the single-channel broadband receiver often intercepts the multiple independent emitter signals. A new algorithm was proposed, in which SSA(singular spectrum analysis)and BSS methods were jointly used. The single-channel signal was firstly changed into pseudo-array signals via SSA method. Then, each source signal was separated via a conventional BSS algorithm. Simulation results indicate that the proposed algorithm can effectively separate the single-channel signals which consist of the signals normally used in electronic reconnaissance, for example, when the single channel signal was mixed by a single-frequency signal and a linear frequency modulation signal, the semblance coefficient between original and separated signals are greater than 0.9 when SNR is 6dB. For the single-channel signal mixed by the different phase shift keying signals, their bit error rate is less than 0.01 after the source separation when SNR is 4dB.

Key words: Signal processing, Singular spectrum analysis, Single-channel blind signal separation, Space electronic reconnaissance