An adaptive beamforming approach using online learning neural network

The paper proposes a new approach for the adaptive beamforming of an antenna array using an online learning based radial basis function (RBF) neural network. The number of hidden layer nodes can be increased or decreased on line, and the function´s centers and widths of the RBF network can be adaptively modified. This network has better generalization performance than that of the conventional K-mean based RBF network. Simulated examples confirm the validity of this method.