Abstract: B1ind signal separation(BSS) is a useful technology which can recover the unknown source signals from the received signals. It has become a new research hot spot of neural network and signal processing. We first introduced the development of BSS and then introduced the BSS's math models which containe linear instant model,linear convolutive model and nonlinear model. We also analysed the algorithms' principal and characteristic of every model and introduced blind signal extraction technology which is closed to BSS. Finally, we pointed out the research direction and wide fore-ground of BSS.

Key words: B1ind signal separation, Independent component analysis, B1ind source separation, Survey