A Fast eigenspace-based beamforming algorithm based on Forward-backward smoothing

The eigenspace-based (ESB) beamformer provides much robustness to pointing errors. However, it couldn´t deal with small sample size which could degrade its performance severely. Forward-backward smoothing technique can improve the performance of beamformers with small sample size. Base on the character that the Forward-backward smoothing matrix is an generalized symmetrical conjugate matrix, The paper deduces an estimation method using real number, reducing the computational cost of estimation by 75%. And then utilizing the real number estimation to the ESB algorithm, the new algorithm improves the performance of the ESB algorithm, as well as reducing the computational cost of weight vector calculating.