A method for identification of optimal minimum number of multi-GNSS tracking stations for ultra-rapid orbit and ERP determination

Abstract

With the rapid increase in the numbers of new generation Global Navigation Satellite Systems (GNSS) satellites, signal frequencies and ground tracking stations, the burden on data processing increases significantly, especially for those real-time or near real-time applications, e.g. generating ultra-rapid satellite orbit and Earth rotation parameters (ERP) products. In order to reduce the number of observations used to estimate the orbit and ERP unknown parameters for better computational efficiency, this study first introduced a parameter called orbit and ERP dilution of precision (OEDOP) factor and a method in “optimally” selecting multi-GNSS tracking stations based on the OEDOP factor is investigated to minimize the data processing burden without significantly sacrificing the accuracy and precision of the satellite orbit and ERP determination. The trade-off between computational efficiency and quality of results is primary focus of this research. The contribution of each tracking station to the precision of the parameter estimates is investigated first, according to the location and multi-GNSS data measurement capacity of the station as well as the length of observations, then those stations that contribute least will be identified and excluded in the estimation system. It aims to use as a fewer number of tracking stations as possible but the degradation in the precision of the solution is still under a desired level. The method was tested using GNSS observations from 409 International GNSS service (IGS) stations over a one-month period. Results showed that when the “degradation” factor of the precision of satellite orbit and ERPs solutions is 5%, 10%, 15% and 20% the accuracy of the satellites orbit and polar motion parameters estimated from an optimal minimum number of stations (in comparison with the results from all stations) reduced about 0.33–9.92 cm and 5.77–41.53 μas respectively, and the accuracy of UT1–UTC reduced 10.63–15.50 μs; while their computational speed was improved by 196%, 332%, 527% and 617% respectively. This suggests that our method is a good trade-off method and an ideal option in cases that rapid solutions are required, e.g. ultra-rapid determination of orbit and ERP using multi-GNSS measurements from global ground tracking stations.