Minimum noise subspace: concepts and applications

Abstract

We present the concepts and some applications of the minimum noise subspace (MNS) technique. The MNS technique has been first introduced as a computationally efficient subspace technique which exploits a minimum number of noise vectors for multichannel blind system identification. It is shown that a noise subspace basis can be obtained in a parallel structure from a set of tuples (combinations) of system outputs that form a properly connected sequence (PCS). The technique of the MNS and particularly the concept of PCS turn out to be powerful tools that can be applied for number of array processing problems. To illustrate the potential of this technique, we present three successful applications related to the problems of blind system identification, source localization, and array calibration respectively.