The TechCrunch article NASA details Deep Space Atomic Clock and other tests launching on SpaceX Falcon Heavy says:

Jill Seubert, Deep Space Navigator for NASA, explained that this is the world’s first ion-based atomic space clock. “It’s about 50 times more stable than the GPS atomic clocks we use,” adding that we currently have to navigate from Earth because the clocks on board spacecraft are really not very good at maintaining time accuracy.

and then later says:

That will enable one-way tracking, when paired with data gathered by an onboard camera, using a signal from Earth to the spacecraft, or from the spacecraft to Earth, but with no round-trip needed. This allows for more efficient tracking across all flights, because you do less time sharing with an existing deep space network. It also enables “self-driving spacecraft,” as Seubert put it, which requires no direction at all from navigators on earth. (emphasis added)

Question: Exactly how does an atomic clock facilitate a “self-driving spacecraft” when paired with an onboard camera?

Is it still necessary to pick up a beacon time signal from Earth? What exactly is the camera for?


  • $\begingroup$ May be the camera is used as a star seeker. But using star seekers was done for decades. $\endgroup$ – Uwe Jun 11 '19 at 14:11

I talked with the two leads of the DSAC in the recent AMA on reddit. Here are my takeaways. First I will point out this great answer, which explains how the clock can be used to simplify navigation with the DSN. I will focus specifically what the camera is for.

There are two types of orbit determination: Absolute navigation and relative navigation. Absolute navigation answers the questions where am I in the solar system in an inertial frame. Relative navigation answers the question where am I relative to some object of interest (usually a another spacecraft). For example, the Dragon spacecraft use a LIDAR relative navigation sensor for proximity operations during docking with the ISS. Cameras have been proposed for replacing lidars in relative navigation

Relative navigation suffers from a problem, It is not fully observable. You can't (in general, there are ways to make it observable) know your absolute orbital state from just relative measurements. Absolute navigation (even if precise down to the meter) generally is not enough for proximity operations; thus the need for relative navigation.

Many spacecraft already use autonomous relative navigation, but could not do on-board absolute determination because of the super strict timing requirements. Now with the DSAC, any spacecraft could have both absolute and relative navigation systems on-board. Once you put trajectory optimization on-board, all you need to do is give spacecraft a destination and it could get there all on its own and do proximity operations there with little to no ground support (except for the DSN broadcasting a timestamp).

Read this DSAC paper and this one to understand the DSAC better. The DSAC leads recommended these papers to me (they also wrote both papers).

  • 1
    $\begingroup$ Nice answer, thank you! I'll give these a read, will take some time to get a copy of the paywalled one though. If possible could you add a link to the AMA? Thanks! $\endgroup$ – uhoh Jun 19 '19 at 1:29
  • 1
    $\begingroup$ I added a link to the AMA. Good catch! Also I switched the pay-walled to a non-pay-walled version. $\endgroup$ – Knudsen Number Jun 19 '19 at 2:02
  • $\begingroup$ excellent, thank you! $\endgroup$ – uhoh Jun 19 '19 at 2:05

With accurate time now available on the spacecraft itself, that can be combined with star tracker and X-ray pulsar (X-Nav) navigation for an accurate fix on position, rotation, and velocity. Previous less accurate spacecraft clocks would require constant recalibration from Earth radio systems with their own atomic clocks.

50 years ago, it was pretty much impossible to do all orbital navigation calculations onboard a spacecraft. Calculations and trajectory determination were done from the ground, and the spacecraft computers were more for autopilot and “flight assist” duties.

  • 1
    $\begingroup$ Your answer doesn't really explain anything. There's no mention of X-ray navigation in the article, and I have not read elsewhere that an atomic lock is necessary when you have X-ray timing from multiple pulsars coming from different directions to compare with each other. Are you just hypothesizing on these two points, or can yo cite a source? Can you explain how a star tracker is used, rather than just repeat that it is used in combination like the article does? Thanks! $\endgroup$ – uhoh Jun 11 '19 at 22:16
  • $\begingroup$ Question: Exactly how does an atomic clock facilitate a “self-driving spacecraft” when paired with an onboard camera? $\endgroup$ – uhoh Jun 11 '19 at 22:19
  • $\begingroup$ You need a reference time to compare the X-ray timings to. Also, for the “one-way” (“self-driving”) tracking, you just use a signal from Earth to gauge relative speed and velocity compared to the Earth. I’ll edit my answer to explain how a star tracker works. $\endgroup$ – CourageousPotato Jun 11 '19 at 22:24
  • $\begingroup$ Can you please cite a source and explains why you need a time reference, and why triangulation by comparing four or more X-ray timings to each other is insufficient? Please ping me when your edits are done and I'll have a look. Thanks! $\endgroup$ – uhoh Jun 11 '19 at 22:31
  • $\begingroup$ See Pulsar clock (mentioned here) $\endgroup$ – uhoh Jun 11 '19 at 22:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.