Neural networks for blind decorrelation of signals

In this paper we analyze and extend a class of adaptive networks for second order blind decorrelation of instantaneous signal mixtures Firstly we compare the performance of the decorrelation neural network employing global knowledge of the adaptive coe cients with a similar structure whose coe cients are adapted via lo cal output connections Through statistical analyses the convergence behaviors and stability bounds for the algorithms step sizes are studied and derived Secondly we analyze the behaviors of locally adaptive multilayer decorrelation networks and quantify their performances for poorly conditioned signal mixtures Thirdly we de rive a robust locally adaptive network structure based on a posteriori output signals that remains stable for any step size value Finally we present an extension of the locally adaptive network for linear phase temporal and spatial whitening of multi channel signals Simulations verify the analyses and indicate the usefulness of the locally adaptive networks for decorrelating signals in space and time accepted for publication in IEEE TRANSACTIONS ON SIGNAL PROCESSING Special Issue on Neural Networks for Signal Processing EDICS Category No SP Portions of this work were supported by the Japanese Frontier Research Program RIKEN Please address correspondence to Scott C Douglas Department of Electrical Engineering Merrill Engi neering Building University of Utah Salt Lake City UT FAX Electronic mail address douglas ee utah edu World Wide Web URL http www elen utah edu douglas On leave from the Warsaw University of Technology Warsaw Poland Permission of the IEEE to publish this abstract separately is granted