Simultaneous DOA estimation based on Kolmogorov's theorem

The design of a new architecture for signal processing, based on the Kolmogorov's theorem (1957), is addressed. This architecture is applied to solve the problem of source separation. Particularly, an adaptive algorithm is proposed to separate simultaneously all the unknown impinging sources on an aperture of sensors. The implemented framework is composed of two different stages: the first one is the inhibition stage, which turns the problem of estimating simultaneous DOAs (directions of arrival) into problems of a single source DOA estimation; and the second one is the optimisation stage which estimates the required parameter in a single signal context easier than the initial one with a multiple signal. A high order rule for learning is described, it improves the behaviour of the system assuring independence of the outputs.<>