Characterization of snow cover using ground penetrating radar for vehicle trafficability – Experiments and modeling

Abstract

For navigational and field testing purposes, it is highly desirable to obtain information about the geometry and material properties of terrains over which vehicles will traverse, preferably in a non-contact manner. The Ground Penetrating Radar (GPR) has emerged as a leading remote-sensing method to probe subsurface properties of terrains; the method detects the difference of the relative permittivity of adjacent layers; the relative permittivity is a function of the density of the material. To the best of our knowledge, no studies using GPR on snow have been conducted for winter mobility/trafficability. In this paper, we used natural and sieved, single- and double-layered snow of which density was varied in the range from 150 to 450 kg/m 3. Given the time-varying GPR signal of snow layers, we were able to determine the depth and dielectric constant ( ε) of each layer using the layer-stripping method. To better interpret the signal from GPR system, which has its limitations, we also conducted modeling and simulation of GPR operating on layered dry snow; given depth and dielectric constant of snow, we simulated the time-history of GPR signal in a layered medium using the Finite Difference Time Domain (FDTD) program GPRMAX as well as the transmission line theory (TLT). For mechanical properties, density is the most commonly used parameter in characterizing the strength of snow. We also obtained hardness of snow as a function of density using Rammsonde. It was found that hardness and relative permittivity increase with density indicating the potential of correlating relative permittivity to hardness; however, the spread of the data is such that future work is needed to quantify statistically this relationship.