ANN-Assisted Iterative Algorithms for Pattern Synthesis of Conformal Arrays

This letter proposes an efficient iterative method assisted by artificial neural network (ANN) for pattern synthesis of arbitrary conformal arrays. Relying on a fast and accurate far-field evaluation solution based on the near-field Euler rotation and two-dimensional (2-D) layer-wise fast Fourier transform (FFT) techniques, an ANN-ANN and an ANN-FFT iterative algorithms are constructed. The theoretical foundation, iterative process, and synthesis results are presented and analyzed. Numerical examples of synthesizing shaped, flat-top and three-dimensional low-sidelobe beams of different conformal arrays are provided to validate the synthesis performance of the proposed algorithms in terms of efficiency and accuracy.