Many modern navigation systems use an inertial navigation system (INS) often integrated with GPS. The largest source of error in the INS is the unknown horizontal components of the gravity vector. Thus, it is critical to know where and at what spacing (resolution) additional gravity data should be collected in order to support INS operation to within specifications. This paper details a method to answer the gravity spacing question. The technique is based on the total error variance equation of least squares collocation and uses a Gaussian covariance function that is valid over the survey grid area. The resulting matrix equation is invertedrnsymbolically and solved for the grid spacing. The gravity variance and correlation distance needed in the method are computed for each one-degree cell using FFT techniques. Tests of this grid spacing model indicate surprisingly little a-priori gravity information is required. In fact, in rugged mountainous terrain, the gravity variance and correlation can be inferred strictly from Digital Terrain Elevation Data (DTED) due to their high correlation with the short-wavelength gravity field. In areas of smoother terrain, longer wavelengths of the gravity field are dominant and small statistically valid samples of gravity were found sufficient to determine the gravity data spacing necessary to support the INS.
展开▼
