Robust beamforming based on steering vector estimation via iterative algorithm

The look direction error can cause the desired signal steering vector mismatch. With this imprecise knowledge of the look direction, the minimum variance distortionless response (MVDR) beamformer suffers considerable performance degradation due to the self-nulling phenomenon of the desired signal. In order to make the MVDR beamformer robust, an iterative algorithm to estimate the desired signal steering vector is proposed. The estimation is initialized by using the presumed steering vector. The estimated steering vector is used as the updated presumed steering vector to iterate the algorithm. The iteration is terminated when the convergence of the error function is attained. The error function is derived from the deterministic maximum likelihood concept. The proposed iterative algorithm is regardless of whether the parameters are appropriately chosen. The effectiveness and comparison of the performance have been shown via a number of simulation results.