Lunar Navigation Without GPS: How Artemis Astronauts Will Find Their Way

When you open a maps app on your phone, you’re the beneficiary of a constellation of satellites, a ground infrastructure that maintains and corrects them, and decades of engineering that has made centimetre-scale positioning essentially free to access. You probably don’t think about how GPS works because you don’t have to. It just works.

That’s not going to be the situation on the Moon. There is no GPS on the Moon. There is no existing navigation infrastructure of any kind in cis-lunar space. When Artemis astronauts eventually stand near the lunar south pole and need to walk a kilometre to a target site, orient themselves after a two-week lunar night, or navigate back to the lander during an EVA, they’ll be doing it without the navigation systems most of us have come to treat as unremarkable background services.

What they will have instead is a set of technologies that are either already proven or in active development. The navigation problem for lunar surface operations is fascinating, and the solutions being developed reveal a lot about how different the Moon’s operating environment is from anything we’re used to thinking about.

Why GPS Doesn’t Work on the Moon

GPS satellites orbit Earth at about 20,200 km altitude. The Moon orbits Earth at an average distance of roughly 384,400 km—about 19 times further than the GPS constellation. GPS signals are directional broadcasts aimed at Earth’s surface, and while the signals technically propagate beyond the GPS orbital altitude, they’re not designed for lunar surface reception and the geometry is entirely wrong for reliable positioning.

You can receive GPS signals from the lunar surface with specialised receivers—research has demonstrated this—but the geometry produces extremely poor dilution of precision (DOP), the technical measure of how well the satellite geometry supports an accurate position fix. On Earth, you typically have multiple satellites in view across the sky at good angles; on the Moon, GPS satellites are all in roughly the same part of the sky (toward Earth), creating terrible geometric coverage. The resulting position uncertainty is on the order of hundreds of metres, not metres—essentially useless for precision EVA navigation.

The Galileo (European), GLONASS (Russian), and BeiDou (Chinese) constellations have similar characteristics. They’re all Earth-orbital systems optimised for Earth-surface positioning.

What the Apollo Astronauts Did

Apollo astronauts solved the navigation problem with a combination of methods that were remarkably effective given the constraints.

The primary tool was simple but carefully calibrated: dead reckoning from the Lunar Module. Astronauts tracked distance walked using odometer readings from the Lunar Roving Vehicle (on J-missions) and time estimates on foot. They maintained a mental model of direction using the Sun’s position and shadows—on the Moon, the Sun’s angle is predictable and consistent over an EVA’s timescale because the Moon rotates so slowly relative to its solar day.

For detailed terrain navigation, the crews used maps produced from Lunar Orbiter photography—high-resolution images from the late 1960s that were used to create topographic charts of the landing zones. Astronauts memorised the local terrain features before landing, allowing them to identify their position relative to craters and surface features during EVAs.

This worked for the Apollo missions because the operations were relatively limited in scope—typically a few kilometres from the lander, for a few hours, with Earth-based support teams who had independent tracking of the crew’s location via telemetry and voice communication. For Artemis, the planned operations are significantly more ambitious, with longer EVAs, more complex terrain in the permanently shadowed regions near the south pole, and requirements for operations that may eventually need to be semi-autonomous.

The LunaNet Architecture

NASA’s answer to lunar navigation infrastructure is called LunaNet—an internet-like framework for lunar communications and navigation services that’s been in development since around 2020. LunaNet isn’t a single system; it’s an architecture specification that defines how multiple assets (orbiters, surface nodes, relay satellites) can work together to provide consistent communications and navigation coverage.

For navigation specifically, LunaNet draws on the concept used for GPS: ranging signals from known orbital positions combined with precisely maintained timing to derive position on the surface. The initial constellation being planned consists of two or three relay satellites in elliptical lunar orbits specifically chosen to provide continuous line-of-sight to the south polar region—the area where Artemis surface operations will occur.

The Lunar Pathfinder mission, developed by ESA with lunar navigation payload contributions, is an early step in this direction. It’s designed partly to demonstrate communications relay capability and to provide data on navigation signal propagation in the lunar environment.

NASA’s own relay satellite plans, part of what’s being called the LCNS (Lunar Communications and Navigation Service), aim to provide positioning accuracy of roughly 50–100 metres from orbit in early configurations, with potential for metre-level accuracy as the constellation expands and ground truth calibration data accumulates.

Terrain-Relative Navigation

Orbital navigation infrastructure is one piece of the picture; terrain-relative navigation is another, and arguably more important for close-range EVA operations.

Terrain-relative navigation (TRN) works by matching real-time sensor data—camera images or lidar scans of the surface—against a pre-built map of the area to determine precise position. It’s essentially asking: “Based on what I can see around me, where am I in the map?”

This technology has already demonstrated impressive results on Mars. The Mars 2020 Perseverance rover used TRN to achieve landing accuracy within 40 metres of its target—a dramatic improvement over previous landing systems. The MOXIE instrument and other hardware choices for Perseverance were made possible by the higher landing precision that TRN enabled.

For lunar surface operations, TRN combined with detailed orbital imagery (the Lunar Reconnaissance Orbiter has been building an extraordinary terrain database of the Moon since 2009) would allow an astronaut’s EVA suit or rover to continuously determine its position relative to the lunar surface to very high accuracy, even without orbital navigation signals. The computational challenge is significant but tractable with current hardware; the key requirement is the high-resolution terrain map to match against, which the LRO data largely provides for the south polar region.

Several teams are working on suit-compatible navigation systems that combine inertial measurement (accelerometers and gyroscopes) with periodic TRN updates to provide continuous position awareness that degrades gracefully if terrain correlation is temporarily unavailable.

The South Pole Navigation Challenge

The south polar region presents specific navigation challenges that make this harder than the Apollo landing zones.

First, the terrain is far more rugged. The south polar region is dominated by heavily cratered highlands, with dramatic relief including the Shackleton crater rim and the complex terrain of sites like Haworth and Nobile craters. Apollo landed in relatively flat mare regions; Artemis will be operating in considerably more challenging topography.

Second, the permanently shadowed regions (PSRs) that are the primary scientific and resource targets are in permanent darkness. No solar-based orientation is possible inside a PSR. Navigation inside these regions requires either a pre-positioned lighting system, lidar-based terrain mapping, or radio-based ranging from beacons outside the PSR—possibly all three.

Third, the communications geometry is different. The south pole is near the edge of the Moon’s nearside/farside boundary, and depending on exact terrain, Earth may not be continuously visible for direct-link communications and tracking. Relay satellites are essential here—not optional.

The Navigation System Artemis III Astronauts Will Actually Use

For the planned Artemis III crewed surface mission, the navigation stack will likely be a layered combination of systems rather than any single technology.

At the base layer: inertial navigation. The EVA suits will have IMUs (inertial measurement units) that continuously track orientation and relative movement. This drifts over time without corrections—an IMU will accumulate position error at roughly 100 metres per hour in a well-designed system—but it provides a continuous position estimate even when other systems are unavailable.

On top of that: radio ranging from the Starship Human Landing System and any surface assets. The lander itself is a known position, and radio ranging can provide a crew position relative to it with reasonable accuracy even without orbital infrastructure.

When orbital navigation signals are available from relay satellites: GPS-like ranging from cis-lunar orbit, providing absolute position on the lunar surface to whatever accuracy the early constellation supports (probably tens of metres for Artemis III).

And for close-range precision: terrain-relative navigation using suit-mounted cameras matched against the LRO terrain database, providing position accuracy of a few metres or better within the pre-mapped EVA area.

Earth-based tracking of astronaut positions through communications signals provides an independent verification layer and allows mission control to maintain a continuous situational picture.

The Longer View

The navigation infrastructure we’re describing for Artemis is, in historical terms, roughly analogous to where Earth satellite navigation was in the 1970s—functional but limited, requiring careful operation within known constraints, and very far from the seamless invisible service that GPS has become for billions of people.

Building out a full lunar navigation constellation to GPS-equivalent utility would require something like eight to twelve satellites in carefully chosen orbits, decades of operational data to characterise the signal environment, and ground truth calibration across the lunar surface as human operations expand. That’s a project that will take as long as building GPS took—measured in decades, not years.

What the Artemis programme is doing right now is building the first elements of that infrastructure while designing operations that work within its current limitations. The astronauts won’t have an iPhone maps experience on the Moon. What they will have—when the systems being developed come together—is enough to get to their targets, complete their objectives, and get back safely. That’s the real engineering challenge, and it’s further along than most people realise.