An improved beamforming algorithm based on feature space

The results of minimum variance (MVDR) algorithm map to the signal subspace to get the weighted coefficients based on feature subspace theory, In the low SNR environment, there will be the main lobe offset, waveform distortion, serious decline in the performance of beamforming. This paper puts forward an improved method for the emergence of this kind of situation. Using the principle of correlation matrix, the covariance matrix of the original algorithm is restructured in order to isolate signal subspace and noise subspace better. Using Unitary transformation, the complex matrix computation transforms to the real matrix computation. While reducing the amount of computation, the algorithm has robust performance in the case of fewer samples. Finally, computer simulation is used to verify the effectiveness of the improved method.