A Robust Nonlinear Beamforming Assisted Receiver for BPSK Signalling

摘要

Nonlinear beamforming designed for wireless communications is investigated. We derive the optimal nonlinear beamforming assisted receiver designed for binary phase shift keying (BPSK) signalling. It is shown that this optimal Bayesian beamformer significantly outperforms the classic linear minimum mean square error (LMMSE) beamformer at the expense of an increased complexity. Hence the achievable user capacity of the wireless system invoking the proposed beamformer is substantially enhanced. In particular, when the angular separation between the desired and interfering signals is below a certain threshold, a linear beamformer will fail while a nonlinear beamformer can still perform adequately. Blockadaptive implementation of the optimal Bayesian beamformer can be realized using a Radial Basis Function network based on the Relevance Vector Machine (RVM) for classification, and a recursive sample-by-sample adaptation is proposed based on an enhanced ?-means clustering aided recursive least squares algorithm.