An adaptive LCMV beamforming algorithm based on dynamic selection of constraints

A novel approach to linearly constrained minimum variance (LCMV) beamforming based on dynamic selection of constraints (DSC) is proposed. The method employs a multiple parallel processors (MPP) framework, where each processor is optimized subject to a particular linear constraint. A selection criterion is employed at the output of the scheme to select the best processor for each time instant. We also present low-complexity stochastic gradient (SG) and recursive least squares (RLS) adaptive algorithms for the efficient implementation of the proposed scheme. Furthermore, a complexity analysis of the proposed and existing schemes in terms of the number of multiplications and additions is carried out. Simulations results with a uniform linear array (ULA) model show that the proposed scheme and algorithms significantly improve the performance of conventional LCMV beamforming.