摘要
经验模态分解算法分解的本征模态函数的个数不能确定且容易出现模态混叠。因此,基于经验模态分解的单通道盲源分离算法存在不稳定性。针对以上问题,选择用变分模态分解算法代替经验模态分解算法,将单通道转换成虚拟多通道,使得欠定的病态问题转换成正定的一般问题。最后,利用独立分量分析算法分离得到源信号。仿真实验说明,该方法可以有效分离出源信号,与经验模态分解方法相比更具稳定性和有效性。
The number of intrinsic modal functions decomposed by the empirical mode decomposition algorithm cannot be determined and modal aliasing is prone to occur.Therefore,there is instability in the single-channel blind source separation algorithm based on empirical mode decomposition.To solve above problems,the variational modal decomposition algorithm is used instead of the empirical modal decomposition algorithm to convert a single channel into a virtual multi-channel,so that the underdetermined ill-posed problem is converted into a positive definite general problem.Finally,the source signal is separated by using an independent component analysis algorithm.The simulation experiment indicates that this method can effectively separate the source signal and is more stable and effective than the empirical mode decomposition method.
作者
吴钟良
王敏
WU Zhong-liang;WANG Min(School of Electronics and Information,Jiangsu University of Science and Technology,Zhenjiang Jiangsu 212003,China)
出处
《通信技术》
2018年第4期774-777,共4页
Communications Technology
关键词
单通道盲源分离
变分模态分解
虚拟多通道
独立分量分析
single channel blind source separation
variational mode decomposition
virtual multi-channel
independent component analysis
