摘要
介绍了基于空间时频分布的循环平稳盲源分离方法,对空间时频分析中交叉项引起的分离能力下降的情况进行了分析,同时将降噪处理应用到算法中,能有效地降低交叉项和噪声的干扰,如选择更有效的时频点进行空间矩阵联合对角化,最终将未知混合信号分离。仿真结果表明,该算法具有较好的收敛性,分离效果明显。
In this paper,the authors introduced the method of cyclostationary blind source separation based on time-frequency distribution,and analyzed decreasing of separation ability which caused by cross-term in analyzing time-frequency.At the same time,the authors applied the noise reduction algorithm,which effectively reduce cross terms of interference and noise,such as selecting the more effective when the frequency on space matrix joint diagonalization,will eventually unknown mixed signal separation.The simulation results showed that the algorithm had good convergence and obvious separation effect.
出处
《科技创新与生产力》
2012年第4期107-109,共3页
Sci-tech Innovation and Productivity
关键词
盲源分离
循环平稳
联合对角化
平滑伪魏格纳维尔分布
blind source separation
cyclostationary
union diagonalization
Smoothing Pseudo Wigner Distribution