A Computationally Efficient Method in Inverse Scattering Using Fourier Series Expansion in Conjunction with FDTD and PSO

A computationally efficient algorithm for solving 1D inverse scattering problems in time domain is proposed. The FDTD and particle swarm optimization (PSO) methods are utilized in conjunction with cosine Fourier series expansion in suggested procedure and the performance is studied for various 1D permittivity profile reconstruction cases. It is shown that the number of unknowns and consequently the amount of computations is reduced considerably while the precision of the solutions would not be affected significantly. Sensitivity analysis of the suggested method to some important parameters of the algorithm is studied briefly, as well.