Signals extraction from their mixture (elements of algorithms and neuro net structures)

Possible accesses and methods are given that enable to get the form of separate signals on the base of measurement information on their mixture. Algorithms and neural network structures realizing this task are based on different concepts that represent features of signal mixture formation, presence of noise at signal registration and also the possibility of noise cancellation for separated signals.