Abstract: A method for dynamic blind signal separation generates initialisation information by processing an immediately preceding data window. This information is input at (21) and used at (22) to initialize orthogonality of data in an immediately following window. Initialised data are decorrelated at (23) with small update angles using a Jacobi technique. Steps (22) and (23) are collectively a second order stage of processing in statistical terms producing orthonormal signals. The orthonormal signals are initialized at (25) and then undergo separation at (27) by ICA with small angle updates using statistics higher than second order to produce separated signals. The method may be implemented in an acquisition phase to separate signals and among them identify desired signals, and a subsequent phase in which only the desired signals are separated.

Abstract: A signal separation method for separation of source signals from a composite signal expresses the composite signal as a series of values of signal amplitude. The source signals have periodicities similar or equal to p. The composite signal is portioned into sections which provide respective rows of a matrix X, in which successive rows represent successive sections. A singular value decomposition of the matrix X is performed to obtain two singular vector matrices U, V and a singular value matrix ?. An independent component analysis is performed on one of the singular vector matrices U, V to obtain an independent component matrix UR2T,R1TV and an associated component matrix R2?V ,U?R1. One of the component matrices UR2T,U?R1 contains estimated separated signal modulation envelopes and the other component matrix R2?V, R1TV contains estimated separated cyclets.