Comparison of multivariate linear regression and neural network algorithms for Ground Penetrating Radar (GPR) estimation of track modulus

Ground penetrating radar (GPR) is a non-destructive and non-invasive technology that has recently been applied to assess the integrity of track substructure, including estimation of track modulus. A single electromagnetic pulse of energy at an appropriate frequency is launched into the ground, and reflections from various subsurface layers are recorded.