A constrained version of fitting pdf algorithm for blind source separation

Abstract

An adaptive multiuser constrained version of a fitting probability density function (p)dAf)' a%@lgorithm for source separation is proposed. The novel algorithm is derived from the multiuser kurtosis (MUK) procedure proposed in the literature. In fact, the proposed approach cay be viewed as a generalization of MUK one. The resulting algorithm is evaluated through computational simulations in the context of a space division multiple access (SDMA) system. The obtained results show that a better performance is attained w.r.t. the unconstrained version.