A neural adaptive beamforming system to reduce interfering signals in mobile communications

This work proposes a neural adaptive beamforming system combining a feedforward artificial neural network with a backpropagation learning algorithm and linear, planar and circular antenna arrays. It intends to reduce interfering signals of the type for multi-path, co-channel interference or near-far field problems in communication systems. The system as a whole maximizes the signal/interference rate (SIR) by the orientation control of the radiation pattern of the arrays. The array individual element excitation allows the creation of a radiation pattern orienting the beam in different directions, pointing out a null (or minimum) for the interfering signal, and a maximum for the desired one.