Analyzing the Impact of Different Pcv Calibration Models on Height Determination Using Gps/Glonass Observations from Asg-Eupos Network

The integration of GPS with GLONASS is very important in satellite-based positioning because it can clearly improve reliability and availability. However, unlike GPS, GLONASS satellites transmit signals at different frequencies. This results in significant difficulties in modeling and ambiguity resolution for integrated GNSS positioning. There are also some difficulties related to the antenna Phase Center Variations (PCV) problem because, as is well known, the PCV is dependent on the received signal frequency dependent. Thus, processing simultaneous observations from different positioning systems, e.g. GPS and GLONASS, we can expect complications resulting from the different structure of signals and differences in satellite constellations.

The ASG-EUPOS multifunctional system for precise satellite positioning is a part of the EUPOS project involving countries of Central and Eastern Europe. The number of its users is increasing rapidly. Currently 31 of 101 reference stations are equipped with GPS/GLONASS receivers and the number is still increasing. The aim of this paper is to study the height solution differences caused by using different PCV calibration models in integrated GPS/GLONASS observation processing. Studies were conducted based on the datasets from the ASG-EUPOS network. Since the study was intended to evaluate the impact on height determination from the users’ point of view, a so-called “commercial” software was chosen for post-processing. The analysis was done in a baseline mode: 3 days of GNSS data collected with three different receivers and antennas were used. For the purposes of research the daily observations were divided into different sessions with a session length of one hour. The results show that switching between relative and absolute PCV models may cause an obvious effect on height determination. This issue is particularly important when mixed GPS/GLONASS observations are post-processed.