An "SVD+Viterbi" algorithm for multiuser adaptive blind equalization of mobile radio channels

Abstract

A fully adaptive algorithm for multiuser blind channel equalization is presented. The algorithm is based on an adaptive matrix singular value decomposition (SVD) for a (virtual) channel identification type operation and the Viterbi algorithm for subsequent symbol detection. Unlike other blind multiuser detection schemes that have appeared in the literature, the present algorithm removes multiple access interference (MAI) by exploiting the finite alphabet property of the input signals together with channel coding redundancy. The latter is believed to be a crucial new ingredient for performance in the context of MAI suppression.