High-Precision SLL Minimization for Uniformly Excited Sparse Arrays Using Grid-less Gradients

Abstract—In this paper, we apply a gradient descent-basedmethod to the synthesis of uniformly excited sparse phasedarrays. The effectiveness of gradient-based optimization in thisdomain is highly dependent on the precision of the Peak SidelobeLevel (PSLL) estimation. Conventional grid-based calculationmethods fail to provide the necessary accuracy, introducingnumerical noise that prevents efficient convergence. To addressthis, we implement a grid-less analytical framework for radi-ation pattern analysis. Our approach not only achieves near-machine precision but also significantly enhancescomputational performance compared to traditional discrete gridsearch. Leveraging this high-precision engine, we develop anupgraded gradient descent strategy capable of synthesizing large-scale arrays. The proposed method effectively manages high-dimensional search spaces and demonstrates superior perfor-mance over heuristic algorithms in terms of both convergencespeed and the depth of sidelobe suppression.