Localizing and identifying underground objects using gravitational measurements

We consider the problem of localizing underground objects and identifying their parameters by measuring the components of the gravity gradient tensor. Applications are oriented towards finding tunnels, pollutants and aquifers, archaeology, oil exploration, verifying compliance with peace treaties (e.g. missile inspection), etc. Close interaction between the physical model and a signal processing approach leads to a source location and parameter estimation problem. We propose several modeling alternatives to solve the non-uniqueness of the inverse problem, though this paper includes only point masses. We use a maximum likelihood procedure to solve the parameter estimation problem. The Cramer-Rao bound is computed and presented for the range and mass of a spheroidal object. Estimation bounds on the location of a single point mass and the resolution for two point masses are studied