Sidelobes reduction and steerable nulling of antenna arrays using neural networks

We present two new techniques for the signal processing of antenna array outputs to produce desired responses either as a complete side lobe canceller or as a steerable interference filter. In the first case, an almost ideal pattern is obtained, However it requires a relatively long training time. However, once the NNs are trained they can be directly used in on-line applications without further need for training. It will be effective in nulling interference from any side-lobe direction while keeping an arbitrary desired response in the main lobe. The second technique necessitates the use of a suitable method for the estimation of the direction of interference signals. Once, these directions are determined the training time time to cancel them will be relatively very short and hence the training can be performed on-line. The presented NN processing techniques overcome many of the limitations of conventional antenna array synthesis and adaptive techniques and are also characterized by reduced cost and simplicity in implementation.