An adaptive cross approximation recompressed degenerate kernel algorithm with radial bases for scattering from PEC objects

A fast algorithm based on hierarchical matrix (H-matrix) is addressed to solve electromagnetic scattering from 3-D electrically large object with arbitrarily shape. According to the admissibility condition of H-matrix framework, the moment-method-based coefficient matrix is divided into near-field and far-field interaction blocks, which are approximated by interpolation degenerate kernel functions. The Gaussian radial basis function (RBF) is adopted to construct the interpolation bases, with a novel interpolation scheme to reduce the interpolation points. The transfer matrices between field and source regions are recompressed with the adaptive cross approximation (ACA) algorithm. Numerical results verify the accuracy and the efficiency of the proposed method.