Adaptive beamforming technique based on neural network

Adaptive beamforming can achieve better SNR by varying the weights of each of the sensors used in the array. The traditional beamforming methods cannot achieve the optimal performance in beamforming because of the mismatch between the assumed array response and true array response. A radial basis function neural network algorithm has been proposed in this paper to solve this problem by turning the processing of calculating weighting of arrays to mapping processing. The simulation results indicate that the proposed method can adapt the weighting according to the direction of source signal automotive, and the SNR can be increased significantly with the DOA mismatch at 2 degrees.