摘要
基于独立分量分析(ICA)的盲源分离(BSS)是一种多源信号分离的优化方法。针对ICA通道数目必需不少于源信号数目的限制条件,提出一种基于频谱识别的重盲源分离(R e-BSS)方法,利用虚拟通道实现通道数目的增加。单通道测量信号的仿真实验和实际信号处理结果表明,该方法简单可行,为盲源分离的进一步工程应用提供了新的思路与方法。
Blind source separation（BSS） based on the independent component analysis（ICA） is a kind of optimal method for multi-source signal separation.For the ICA issue,the number of the test channel not being less than that of the signal source,a new method of repeated blind source separation（Re-BSS） based on power spectral analysis is proposed.This method can increase the number of test channel by means of setting a virtual signal channel.The signal simulation of one single channel and the signal processing of a practical measurement show that the proposed Re-BSS method is simple and feasible for operation and provides a new way for the further engineering application of the BSS.
出处
《振动工程学报》
EI
CSCD
北大核心
2010年第5期508-513,共6页
Journal of Vibration Engineering
基金
国家自然科学基金项目(50975202)
国家863项目基金资助(2007AA04Z414)
关键词
独立分量分析
频谱分析
盲源分离
重盲源分离
independent component analysis
frequency spectrum analysis
blind source separation
re-blind source separation
