New beamforming method based on radial- basis function neural network processing in SαSG distribution noise environments

Abstract

This paper considers the beamforming problem with radial-basis function network in alpha stable noise environment. In the new noise environment, a novel training method is proposed based on covariation. Comparing the output of the network with the analytical solution, it is found that they are very consistent. Then, it is reasonable to perform beamforming by using radial-basis function network.