Developing Realistic FDTD GPR Antenna Surrogates by Means of Particle Swarm Optimization

The antenna is the most important part of a ground-penetrating radar (GPR) system and defines employed electromagnetic pulse how it transferred to ground. It crucial account for these coupling effects in numerical simulations implement realistic models, e.g., full-waveform inversion (FWI). We present method developing adapting 3-D finite-difference time-domain (FDTD) models GPR antennas, complete with electric components, dielectric material properties, feed details. exemplify this commercially available, shielded 400 MHz antenna, model which was set up by fitting synthetic data an experimental signal reflected at metal plate air. For FWI, we used particle swarm optimization (PSO) algorithm because fit parameters show complex individual on waveform. resulting then validated against measured air, water, near field antenna. Overall, reproduce validation very accurately. Signals objects placed change shape frequency content radiated wavelet varying subsurface properties are emulated correctly.