StarFOX Update #1: Publishing lunar navigation results.

This past week, we presented initial results of applying ARTMS navigation in lunar orbit at the ION International Technical Meeting in Long Beach.

ARTMS, or the Absolute and Relative Trajectory Measurement System, is the multi-spacecraft autonomous navigation architecture for StarFOX. StarFOX will test ARTMS in low Earth orbit, but ARTMS itself can be generalised to any orbit – and one scenario of particular interest is lunar/cis-lunar exploration. NASA (via their Artemis program, no relation) and many other governments, commercial companies and research institutions are making a beeline for the Moon as humanity’s first off-world outpost. Supporting that effort requires new methods for navigating in space.

For example, while you can technically lock onto GPS satellites from the Moon if you try really hard, it requires specialised antenna and has limited accuracy. Another traditional solution is relying on Earth-based radio antenna, but that isn’t scalable to lots of spacecraft and leads to worse responsiveness. Instead, it’s often preferable to have navigation be autonomous and self-contained within your system – and that’s where ARTMS shines, thanks to its camera-based approach.

If we can successfully navigate groups of spacecraft in lunar environments, it lets us apply them to many interesting objectives such as lunar communication networks, ground-penetrating radar surveys, or monitoring for Deep Space Gateway operations.

Nevertheless, applying ARTMS in lunar orbit means means overcoming some challenges. These include:

• Very different orbital behaviour compared to StarFOX’s low Earth orbit. Dynamic speeds are generally much lower due to the weaker central gravity field, and the Moon has lots of oddly-shaped mass concentrations that pull on the spacecraft in strange ways. Third-body gravity effects from the Earth and Sun are also very pronounced.
• A need to estimate other system parameters in addition to its orbits. These include estimating the offsets between different spacecraft clocks (to make sure measurements across different spacecraft remain synchronised) and ballistic coefficients (to accurately model how the sun’s radiation pressure affects spacecraft trajectories).
• Challenging visual tracking scenarios. During the StarFOX experiment, each spacecraft is close together, so it’s relatively straightforward for them to take photos of each other. However, in lunar scenarios of interest, spacecraft might be much further apart. Cameras therefore have to actively follow targets to keep them in view.

At SLAB, we’ve done our best to address these questions, with promising results so far! ARTMS has been successfully leveraged to navigate multiple spacecraft in five core lunar scenarios:

1. A swarm in low lunar orbit in a “passive safety ellipse” formation (i.e. with significant relative motion)
2. A swarm in low lunar orbit in an “in-train” formation (i.e. with very little relative motion)
3. A swarm in a medium “frozen” lunar orbit (i.e. a specially-designed elliptical orbit)
4. A swarm in a high “near-rectilinear halo orbit” (i.e. a very elongated and large orbit)
5. A “constellation” in frozen lunar orbit in four orbit planes (i.e. with LOTS of relative motion)

Cases 1-3 are fairly straightforward applications of ARTMS, but Cases 4 and 5 present unique challenges. Case 4 observes very high variations in spacecraft velocity and behaviour because the halo orbit is very elongated. It closest approach to the lunar surface is very low and fast (a few hundred kilometers), and its farthest approach from the lunar orbit is very high and slow (70,000 kilometers). It’s challenging to robustly estimate spacecraft position and velocity because of the resulting dynamical variations. The halo orbit is particularly important, however, because it’s the planned orbit of NASA’s Gateway space station.

The orbit path of the swarm for the NRHO. As can see, it’s a bit of a mess.

Case 5 is difficult because it’s a so-called “constellation”, in which the different spacecraft are separated by thousands of kilometers (instead of the tens or hundreds for for StarFOX). Each spacecraft possesses a completely different orbit, which means a) some mathematical assumptions within ARTMS are no longer valid, and b) it’s hard to visually keep targets within each camera’s field of view. It’s again an important case, though, because constellations provide much wider, more consistent coverage of the lunar environment. This is obviously useful if you’re trying to run a communications network or survey the lunar surface.

An example of a four-spacecraft constellation orbit.

The nice thing is that, with sensible adjustments, ARTMS is able to deal with these scenarios. Much more work would have to be done to fully address all of our operational concerns, but during initial investigations, ARTMS successfully performs angles-only navigation in lunar environments. Here are some plots of absolute and relative navigation results for Cases 4 and 5. You’ll notice a lot of periodic behaviour due to orbit eccentricity, but ultimately, the state estimates converge!

There are some interesting trends to observe in these results (and lots of future work to be done), but I’ll leave that discussion for the full paper if you’re interested!

Special thanks Keidai Iiyama for lending his lunar navigation expertise to ARTMS.