Constrained NLMS Algorithm with Variable Coefficient Reusing

In general, algorithms for adaptive arrays intend to emphasize desired signals while discriminating against interference signals. In this paper, a novel constrained normalized least mean-squares algorithm with improved robustness against noise is proposed. The devised algorithm requires a low computational burden and takes advantage of a variable reusing coefficient technique, in order to enhance asymptotic performance maintaining good convergence behaviour. Simulations reveal that the estimated weights of the proposed algorithm lead to an efficient beam pattern.