GPS-Free Navigation: NASA's FALCON Uses Debris as Landmarks

What happened

NASA says one of the four small spacecraft in its Starling CubeSat swarm has worked out its own orbit without GPS and without a running stream of instructions from the ground. The experiment, called FALCON (Fast Autonomous Lost-in-space Catalog-based Optical Navigation), had the spacecraft photograph other satellites and pieces of orbital debris, identify them, and use them as reference points.12

The method relies on an onboard catalogue of about 20,000 tracked objects. When the spacecraft's cameras captured an object, the system matched it against the catalogue, looked up where that object should be at that moment, and treated those predicted positions as moving landmarks. From several of those fixes, it inferred its own orbit.2 In a separate part of the experiment, the system also refined the tracks of more than 200 catalogued objects.2

NASA describes the self-orbit determination as a first for optical navigation that works from a spacecraft's position relative to other objects in space.2 The agency is not claiming to be the first to navigate without GPS, or the first to steer by camera. Both have been done before.2 What is new is the reference frame: using artificial objects, many of them junk, in place of GPS signals or natural celestial bodies.

A software story riding on existing hardware

The most interesting part of FALCON is how little new hardware it needed. The experiment combined Era-Core, a software package from the company EraDrive, with cameras the Starling spacecraft were already carrying.2 Born To Engineer says these were Starling's star-tracker cameras, sensors normally used to keep track of a spacecraft's orientation, not its location.3

That makes FALCON as much a software and tooling result as a spaceflight one. If an off-the-shelf navigation stack can turn standard attitude sensors into a position-finding instrument, satellite builders could add an independent navigation capability through a software load rather than a new payload. On a CubeSat, where mass, power and volume are tight, that matters. It also points to a growing commercial market for flight software sold as a reusable product, separate from the spacecraft it runs on.

Publicly available accounts say little about EraDrive itself, including how it was founded or how the software is licensed. Claims about the company's background beyond its role supplying Era-Core should be treated cautiously until more details come out.

Why it matters beyond Earth orbit

The coverage agrees that the main payoff lies farther from home. Digital Journal argues that as people and machines push deeper into space, knowing exactly where a spacecraft is becomes as important as planning where it goes.1 It points to the infrastructure that could build up around the Moon, such as communications relays, science orbiters, autonomous landers and robotic support craft, all of which would need reliable position fixes at all times.1 Space Daily likewise frames FALCON as a step toward navigation near the Moon and Mars.2

There is a catch. Low Earth orbit is crowded, which is exactly why a catalogue of 20,000 objects gives a camera so many landmarks to work with. Cislunar space and Mars orbit have far fewer artificial objects. In the near term, the technique looks best suited to busy regions. It could also serve as a backup if GPS is degraded, jammed or unavailable. Wider use elsewhere would depend on how much traffic builds up there. That is an inference, but it follows from how the system works.

The approach also has an appealing side effect. Debris is usually treated purely as a hazard. FALCON treats it as a navigational resource, and while doing so it improves the orbital data for the objects it observes.2 A fleet of spacecraft running similar software could, in principle, help with space situational awareness as a by-product of normal operations.

Where the accounts differ

The reports roughly agree on the substance but differ on timing and tone. Space Daily dates NASA's announcement to 17 August 2026.2 Born To Engineer says the results became public in early September 2026, citing SciTechDaily's coverage.3 ScienceDaily carried the story in its space exploration headlines during the same period.4 The gap more likely reflects when different outlets picked up the news than any change in the findings.

Born To Engineer adds that the results went beyond what the team had expected and compares FALCON to sailors steering by Polaris and Sirius.3 Digital Journal focuses on autonomy, placing FALCON within NASA's broader push for systems that can act without constant human input.1

The takeaway

FALCON is a modest flight experiment with an outsized implication: an existing sensor plus the right software can give a spacecraft a new, GPS-independent way to know where it is. The sailor analogy is apt, but the more practical lesson for engineers is that the hard part was solved in code. If that holds up across more missions, navigation could become something spacecraft operators upgrade through software rather than hardware.