GPS, GLONASS, and Galileo orbit geometry variations caused by general relativity focusing on Galileo in eccentric orbits

Abstract

Three main effects from general relativity (GR) may change the geometry and orientation of artificial earth satellite orbits, i.e., the Schwarzschild, Lense–Thirring, and De Sitter effects. So far, the verification of GR effects was mainly based on the observations of changes in the orientation of satellite orbital planes. We directly observe changes of the satellite orbit geometry caused by GR represented by the semimajor axis and eccentricity. We measure the variations of orbit size and shape of GPS, GLONASS, and Galileo satellites in circular and eccentric orbits and compare the results to the theoretical effects using three years of real GNSS data. We derive a solution that assumes the GR to be true, and a second solution, in which the post-Newtonian parameters are estimated, thus, allowing satellites to find their best spacetime curvature. For eccentric Galileo, GR changes the orbital shape and size in perigee in such a way that the orbit becomes smaller but more circular. In the apogee, the semimajor axis decreases but eccentricity increases, and thus, the orbit becomes more eccentric. Hence, the orbital size variabilities for eccentric orbits are greatly compensated by the orbital shape changes, and thus the total effect of satellite height change is much smaller than the effects for the size and shape of the orbit, individually. The mean semimajor axis offset based on all GPS, GLONASS, and Galileo satellites is − 17.41 ± 2.90 mm, which gives a relative error of 0.36% with respect to the theoretical value.

Characterization of Galileo yaw attitude on tidal loading and range bias in SLR-based orbit validation

“…On the other hand, we use the official Bernese version 5.2, which only use the Schwarzschild term conventionalized by IERS (1996) for the general relativistic correction in the orbit integration. However, the order of magnitude in acceleration caused by the Schwarzschild effect is 10 and 100 times bigger than the De Sitter and the Lense-Thirring term, respectively, roughly at 10 -11 and 10 -12 m/s 2 level (Sośnica et al 2022). Such order of magnitude in acceleration may only cause sub-mm to 1 mm errors in GNSS orbit.…”