A new blind separation approach based on blind localization
Qingwen Liu, Wang Zeng-fu · 2002
In this paper, a new approach for blind source separation (BSS) based on blind localization is presented, which makes use of the inherent relationship between separation and localization. The approach incorporates the advantages of classical algorithms of BSS and localization. It can not only determine the number of signal sources, which is difficult to deal with by traditional separation algorithms, but also reduce the adaptive learning time. The efficiency of the approach is illustrated by simulation.