Blind separators
Blind source separation

Before 1985, antenna processing techniques were essentially of order 2, and required a priori information on sources and/or the sensor network. From this date onwards, the first self-taught source separation techniques appeared, making it possible to dispense with this a priori information, replacing it with the assumption of source independence.

Various approaches were then developed: approaches based on the minimization of crossed moments of nonlinear functions of the outputs [24][28], approaches minimizing crossed cumulants, then approaches by maximizing contrasts [29], maximum likelihood approaches when the noise distribution is known [30]