Modified implementations of RLS digital beamformer

Some modifications of the conventional recursive least square (RLS) beamformer are presented. The proposed modifications allow us both to reduce the computational complexity of RLS beamformer and to increase its numerical performance. The proposed modifications are based upon the application of vector operations instead of matrix operations and triangular decomposition. As shown, the proposed implementation outperforms the conventional RLS beamformer in computational complexity as well as in robustness against the computational errors.