On vectorial integer bootstrapping implementations in the estimation of satellite orbits and clocks based on small global networks

Sebastian Strasser,Torsten Mayer-Gürr,Lotfi Massarweh

doi:10.1016/j.asr.2021.09.023

Sebastian Strasser, Torsten Mayer-Gürr + Show 1 more

Open Access

https://doi.org/10.1016/j.asr.2021.09.023

Copy DOI

Abstract

The estimation of satellite orbits and clocks plays a central role in different Global Navigation Satellite System (GNSS) domains, e.g. precise positioning or time transfer. Such products can be computed in the process of Orbit Determination and Time Synchronization (ODTS), which relies on a network of ground-based stations, well distributed around the globe. The mm-level precision of carrier-phase measurements is exploited in this network estimation following a correct resolution of their ambiguities. For several stations, satellites and/or signals, thousands of ambiguities have to be processed, which means having to deal with high-dimensional ambiguity resolution (HDAR) problems. In this research work, we firstly account for the impact of ambiguity resolution in a varying network size, based on GPS-only, Galileo-only and GPS+Galileo configurations. Using 25 or more stations, the accuracy (1D RMS orbital errors) of fixed solutions reaches a plateau at 1–2 cm. Hence, we focus on a small global network of 14 stations, where the model strength decreases, so does the reliability of the ambiguity fixing process and advantages over a float solution might become less evident. In order to allow reliable HDAR, two implementations of the Vectorial Integer Bootstrapping estimator are presented and evaluated with respect to their scalar counterpart. Finally, it is shown how the proposed fixing processes are more robust, still very efficient, and on certain days they provide a large improvement to satellite products. The orbital results are ultimately validated by considering the satellite midnight discontinuity errors over a 3-month period in 2019.

Highlights

The generation of satellite orbit and clock information represents an essential element for any Regional/Global Navigation Satellite System (RNSS/GNSS) and is generally based on the use of code and phase observations from a ground-based network of station receivers
We start with the Galileo-only processing, where up to 850 ambiguities need to be fixed for each day
All three solutions can largely benefit from the decorrelating Ztransformation that has been previously applied with Least-squares AMBiguity Decorrelation Adjustment (LAMBDA)

Summary

Introduction

The generation of satellite orbit and clock information represents an essential element for any Regional/Global Navigation Satellite System (RNSS/GNSS) and is generally based on the use of code and phase observations from a ground-based network of station receivers. The so-called process of Orbit Determination and Time Synchronization (ODTS) makes the best use of state-of-the-art knowledge on orbital dynamics for all tracked satellites (Montenbruck and Steigenberger, 2020). This a priori information can be improved by means of an accurate functional and stochastic modeling of such measurements, ideally consistent with the models later adopted on the user side. The contribution of carrier-phase measurements, very precise but ambiguous, is a key aspect in the network estimation, whether a global or a regional one.

Results

Discussion

Conclusion