Antarctica is a key region for the development of the climate on Earth. Global Navigation Satellite Systems (GNSS) help to better understand the associated geodynamic processes by deriving precise deformation rates of the solid Earth using repeated or continuous measurements. Besides the determination of plate tectonic movements, the determination of glacial isostatic adjustment (GIA) is one of the most important applications of GNSS in polar regions. GIA describes the response of the solid Earth to changing ice-loads. It manifests itself in a deformation measurable at the Earth's surface, which is mainly caused by redistributions of the viscous mantle material within the Earth. The ice-mass balances determined by satellite gravimetry, which serves as an input variable for climate modelling, are largely affected by errors due to the uncertainties of the GIA-induced mass redistributions. Therefore, GNSS results are very important for the validation of GIA models and studies based on them. In this thesis a consistent processing of all GNSS data was performed which were measured on bedrock in Antarctica. The data were made available within the international cooperation GIANT-REGAIN (Geodynamics In ANTarctica based on REprocessing GNSS dAta INitiative). In turn, the results of this work are a contribution to this project as well. A major problem of previous GNSS studies in Antarctica has been the limited coverage, resulting from either a regional analysis or a limited selection of GNSS sites for investigations on a continental or global scale. Moreover, some important regions were only very rarely considered, such as the Amundsen Sea embayment which is characterized by an extreme ice-mass loss. Various GNSS studies accomplished so far used different processing strategies, input models and reference frames, so that inferred rates cannot be compared directly. Therefore, in most cases a validation of the GIA models is only possible to a limited extent. Now, with the joint processing of more than 250 GNSS sites for the period from 1995 to 2017 the previous limitations could be circumvented. Deformation rates could be determined for almost all sites, which are derived from a homogeneous analysis and are, therefore, directly comparable and interpretable. Besides the processing of GNSS data another focus lies on the treatment of associated metadata. Their correct or insufficient acquisition can have a significant influence on the derived deformation rates. By setting up a data management system including various graphical interfaces the data handling has been made significantly more efficient. In addition, many errors were detected and could be corrected to a great extent. Further aspects of the investigations include the optimization of the geodetic datum definition by adjusting the fiducial site selection, the detection of outliers and jumps in the time series for a reliable trend estimation, and the handling of obvious problems of some sites. Erroneous metadata and ice deposits within the antenna are among the most problematic effects. In this context, it was also shown that automated methods for the detection of outliers and jumps as well as robust methods to mitigate or eliminate these anomalies provide very good results for many sites. However, there still exist several sites with special characteristics where manual revisions are strongly recommended for. Especially the ice deposits within the antennas, which have not been investigated in detail yet, cannot be adequately considered by these methods. Furthermore, realistic measures for the uncertainties of the GNSS deformation rates were derived by a careful accuracy estimation. The deformation rates determined with GNSS in Antarctica were analysed in a geodynamic context. In East Antarctica, the vertical deformation rates are very small with only a few millimeters per year and a frequently changing sign. In contrast, the vertical deformation rates in West Antarctica are much higher and can reach several millimeters per year. The extreme uplift rates in the area of the Amundsen Sea embayment play a special role and were investigated in more detail within this thesis. There, the GNSS rates reach values of up to 62mm/a and, reduced by the effect of recent ice-mass changes, of up to 45mm/a. They represent the largest measured uplift rates due to glacially induced deformations worldwide. With increasing distance to the large glaciers of this region, the rates decrease rapidly, resulting in large gradients. Nearly all GIA models underestimate the GNSS-derived uplift rates by almost an order of magnitude in that area. This is most likely due to the combination of a special rheology (small thickness of the lithosphere and low viscosity of the asthenosphere) and an extreme ice-mass loss during the last decades. As a result, more recent events in the ice-load history dominate the present-day rates, which is why the classical separation of immediate (elastic) deformations and those that persist over millennia does not seem to work in this case. A minor effect of tectonic processes and especially volcanism cannot be excluded but has most likely no significant influence. The horizontal deformation rates in Antarctica mainly reflect plate tectonic motion. After deducting the proportional motion of the Antarctic Plate, the horizontal rates are very small, which also applies to the relative velocities. Therefore, the Antarctic Plate can be considered as very stable overall. Only between the Antarctic Peninsula and the South Shetland Islands as well as in the Amundsen Sea embayment larger horizontal rates can be found, which are caused by a separate lithospheric microplate (Shetland Plate) or by the considerable GIA effect in the Amundsen Sea embayment, respectively.