To determine altitudes in regard to gravity with high resolution based on GPS, regional geoid models are necessary. Up to now they are mostly not available at a practical implementation. The outcome of this is the direct need to create autonomous regional geoid models with high accuracy.

There are three different approaches:

1. exclusive use of terrestrial data,
2. exclusive use of satellite data as well as
3. combined use of terrestrial and satellite data.

The model corresponding to 3) is the optimal solution. Depending on the spatial resolution and distribution of the included gravity values, a precise geoid model of high degree and order and therefore high resolution will be received. A disadvantage of 2) is that only the long-wave and not the short-wave parts of the gravity field will be considered, vice versa for option 1).

As the measurement of a dense gravity field grid is in general very complex, a regional geoid model will be generated by using satellite observations. In addition, terrestrial gravity data can be added successively as far as possible to increase the density. While computing geoid models there should be paid attention to the temporal change of the gravity field. Also the quality and the quantity, respectively the density of the observations are changing frequently so that geoid recalculations are consistently necessary to provide an actual gravity reference surface. But there exist also model approaches where a recalculation can be avoided to some extent.

The project RegGRAV tends to the development of a computer program system for the generation of regional gravity and gravitational fields out of heterogeneous data. This data must be preprocessed and conveyed to an algorithm for the computation of a geoid.