Stochastic-deterministic and sensitivity analysis of the transient field generated by GPR dipole antenna and transmitted into a lossy ground

Abstract

The paper deals with a stochastic-deterministic modeling and sensitivity analysis of the transient field generated by a ground penetrating radar (GPR) dipole antenna and transmitted into a lossy ground in time domain. Uncertainties are inherent to the environmental and geometrical parameters pertaining to GPR applications. Such settings include dielectric properties of a soil, the terrain inhomogeneity, surface roughness, uncertainty in detected depth of buried object etc. The stochastic modelling is carried out by means of Lagrange's stochastic collocation (SC) method while deterministic model is based on the Time Domain Integral Equation (TDIE) approach and related numerical solution via the space-time variant of Galerkin Bubnov's Indirect Boundary Element Method (GB-IBEM). The presented model provides a mean to rank the input parameters from the most to least influential one and to obtain statistical moments for the transient field values. The paper also aims to combine the time domain techniques with the stochastic method to provide a more complete description of GPR antenna generated fields.