Low-Sidelobe Antenna Beamforming Via Stochastic Optimization

We describe a novel iterative methodology for computing a set of low-sidelobe beamforming weights for an airborne, electronically-steered phased-array radar using an in-flight stochastic optimization routine performed over a number of coherent processing intervals (CPIs). The proposed approach is notable in that it does not rely upon a good antenna calibration and only requires digitization of the radar's sum beam. By observing the radar ground clutter, the algorithm iteratively adjusts the beamformer. Furthermore, it is computationally inexpensive and scales favorably to radars comprising very large numbers of antenna elements and requiring extremely low sidelobes.