Decoupled and coupled moons' ephemerides estimation strategies -- Application to the JUICE mission

TL;DR

This paper compares decoupled and coupled strategies for estimating the ephemerides of moons, demonstrating that coupled methods can significantly reduce uncertainties and better account for dynamical interactions, especially for the Galilean moons during the JUICE mission.

Contribution

It provides an explicit formulation for the coupled estimation approach and compares its performance with the decoupled method using simulated data for the JUICE mission.

Findings

Coupled approach reduces radial uncertainties by over an order of magnitude.

Decoupled approach yields slightly lower tangential position errors.

Coupled method is more sensitive to moons' dynamical interactions.

Abstract

When reconstructing natural satellites' ephemerides from space missions' tracking data, the dynamics of the spacecraft and natural bodies are often solved for separately, in a decoupled manner. Alternatively, the ephemeris generation and spacecraft orbit determination can be performed concurrently. This method directly maps the available data set to the estimated parameters' covariances while fully accounting for all dynamical couplings. It thus provides a statistically consistent solution to the estimation problem, whereas this is not directly ensured with the decoupled strategy. For the Galilean moons in particular, the JUICE mission provides a unique opportunity for ephemerides improvement. For such a dynamically coupled problem, choosing between the two strategies will be influential. This paper provides a detailed, explicit formulation for the coupled approach, before comparing the performances of the two state estimation methods for the JUICE test case. To this end, we used both decoupled and coupled models on simulated JUICE radiometric data. We compared the resulting covariances for the Galilean moons' states, and showed that the decoupled approach yields slightly lower formal errors for the moons' tangential positions. On the other hand, the coupled model can reduce the state uncertainties by more than one order of magnitude in the radial direction. It also proved more sensitive to the dynamical coupling between Io, Europa and Ganymede, allowing the solutions for the first two moons to fully benefit from JUICE orbital phase around Ganymede. However, many issues remain to be solved before a concurrent estimation strategy can be successfully applied to reconstruct the moons' dynamics over long timescales. Nonetheless, our analysis highlights promising ephemerides improvements and thus motivates future efforts to reach a coupled state solution for the Galilean moons.