Broadband beamforming for joint interference cancellation and turbo equalization

This paper considers the turbo equalization of trellis-coded modulated (TCM) broadband wireless signals that are affected by both intersymbol interference (ISI) due to multipath propagation and cochannel interference (CCI) due to the presence of adjacent users. Long channel dispersion causing both impediments to be severe makes the direct application of trellis-based turbo equalizers impossible, particularly for high-order signal modulations. For this reason, we present two computationally feasible space-time turbo receiver architectures employing linear antenna arrays and broadband beamformers. The first receiver uses a beamformer for joint rejection of interfering signals and shortening of the channel corresponding to the desired signal so that a scalar trellis-based turbo equalizer can be applied to its output. At each beamformer branch, the composite effect of the CCI and the white channel noise is viewed as colored noise. Because trellis search algorithms are limited to cases where the observation noise is white, this imposes quadratic noise whitening constraints on the beamformer design that is solved by a Lagrangian relaxation approach. The second receiver architecture removes the dependence on trellis search techniques for equalization by implementing the turbo equalizer directly with a soft-information-aided broadband beamformer at its front end and a soft-input soft-output (SISO) decoder at its back end. We outline the design considerations associated with each receiver and present bit error rate (BER) simulations for the turbo equalization of 8-phase-shift keying (PSK) TCM signals.