Dual-level compressive sensing method for direct solution of wideband and wide-angle electromagnetic scattering problems

In this paper, a dual-level compressive sensing (CS) method is proposed to efficiently solve wideband and wide-angle electromagnetic scattering problems by accelerating the ultra-wideband characteristic basis function method (UCBFM) on two levels. At the first level, a multiple-measurement-vectors (MMV) CS model is established, aiming to reconstruct the entire set of ultra-wideband characteristic basis functions (UCBFs) from a limited number of measurements. As a result, not only is the computation time of UCBFs reduced, but also the number of UCBFs is decreased. At the second level, another MMV-CS model is constructed to solve the induced current at each frequency point, through which not only can the measurement matrix be constructed by filling the impedance matrix with just a few rows, but also the complicated construction process of the reduced matrix is avoided. Numerical simulation results show that the computational efficiency is significantly improved by the proposed method.