A beamforming method in UWB pulse array based on neural network

In this paper, the principles and algorithm of array beamforming based on a realistic ultrawide-band (carrier free) pulse signal model is discussed firstly. Then, a UWB pulse array beamforming method, based on a radial basis function (RBF) neural network, is presented. It can compensate the propagation delays accurately in each channel and form a narrow directivity beam in a desired direction. Compared with the method of employing FIR filters for delay steering in wideband beamforming, the RBF approach has the advantage in achieving arbitrary array response. The simulation results show that the beam formed by RBF has high main gain and very low sidelobe, in addition to rapid and accurate processing due to the high parallelism and mapping ability of the neural network.