An EM-based beamforming algorithm for WCDMA downlink signals

Multipath propagation causes intersymbol interference and results in a loss of orthogonality between the pilot channel and the user signals in wideband code-division multiple-access (WCDMA) systems. Consequently, an adaptive beamformer using only the pilot signal as the training sequence may not provide adequate error-rate performance. Although a conventional maximum likelihood (ML) beamformer gives improved performance, its computational complexity is relatively high. In this paper, we present a reduced-complexity ML beamformer based on the expectation-maximization (EM) algorithm that utilizes information about the unknown user signals. For comparison purposes, we describe another reduced-complexity ML beamformer based on a Gaussian model for base station self-interference. These beamformers improve the performance of the pilot-only beamformer as demonstrated in a computer simulation example.