A compressed sensing-based approach for null steering in partially adaptive planar arrays using a reduced number of adjustable array elements

This paper proposes two methods for multi-wide null steering in the beampattern of preconfigured partially adaptive planar array antennas using convex optimization and sparse recovery theory. The array elements are split into two categories. One of which is nonadjustable or has fixed excitations, and the other is adjustable. The weights of the adjustable part are perturbed to form wide nulls at the interference directions. In the first proposed method, all the adaptive array elements are used to steer the nulls. Here, convex optimization is used to obtain the required weight perturbations of all the adaptive elements. In addition to the constraints that are needed to impose the nulls, constraints were added to mitigate the pointing error and the lower peak-to-sidelobe ratio (PSLR) problems. In the second proposed method, the optimization problem is cast as a sparse recovery problem to reduce the number of elements to be perturbed from the set of adjustable elements only. The optimization problem is then relaxed to achieve a convex optimization problem and solved iteratively to cut down the number of radiating elements which needs to be perturbed from the set of adjustable elements. Simulations were conducted for different sizes of planar arrays up to a few hundred antennas and different configurations of adjustable elements. The results showed that changing the excitations of only 20-35% of the adjustable array elements is sufficient to steer the nulls without any pointing error or decrease in the PSLR. Also, the interference rejection ratio (IRR) achieved by the proposed methods is better than the minimum mean square error (MMSE) analytical solution.