I have read that sometimes spacecraft needs to correct its orbit but my question is how do they know thier orientation and position in the empty space?

  • 5
    $\begingroup$ (One of) the answer(s) is literally your user name. $\endgroup$ Dec 28, 2021 at 18:49
  • 2
    $\begingroup$ @JörgWMittag Exactly. $\endgroup$ Dec 28, 2021 at 18:56

2 Answers 2


Orientation is easy; it's in your user name. Star trackers are widely used to sense orientation. Rate gyros are often used to track orientation between star tracker readings. There is a potential issue with using rate gyros, which is that they measure rotation rate rather than orientation. Since the measurements aren't perfect, the estimated orientation drifts from the true orientation. This drift isn't a big problem if star tracker updates come in often enough. In addition to star trackers, sun sensors and horizon sensors can be be used to detect orientation.

Position is tougher. Vehicles in low to mid Earth orbit can use GPS to estimate their state and accelerometers to update translational state between GPS updates. The aforementioned issue with orientation using rate gyros becomes more serious with regard to position. Integrating accelerometer readings would yield velocity, but there's a problem here: Accelerometers do not sense acceleration due to gravity. Acceleration due to gravity needs to be estimated based on estimated position, and position estimates result from integrating estimated velocity.

Many vehicles beyond low Earth orbit simply don't know where they are.

  • $\begingroup$ Do you mean the vehicles don't know (and don't care) but we do know (and do care!), or that we have no idea where most vehicles are? I would think that someone has a good idea of where most spacecraft are located to within reasonable accuracy. $\endgroup$
    – JohnHoltz
    Dec 28, 2021 at 19:50
  • $\begingroup$ Thanks for the answer but do you mean that spacecraft actually track locations of star? Because i think even if they move in deep space the location of stars will barely change so how do they know how far they have travelled? $\endgroup$ Dec 29, 2021 at 3:38
  • 2
    $\begingroup$ @JohnHoltz People on Earth do know where their spacecraft are. Knowing that is one of the key purposes of NASA's Deep Space Network, and similar services by other space agencies. It's the spacecraft that don't know where they are. The spacecraft do know where they are pointing. $\endgroup$ Dec 29, 2021 at 6:28
  • 1
    $\begingroup$ @StarsTracker Google your name, or rather, google "star tracker". The constellations look very much the same across the expanse of the solar system. This makes it easy for a spacecraft to autonomously know where it is pointing. Autonomously knowing where it is as much more challenging problem, one that hasn't been fully solved yet. $\endgroup$ Dec 29, 2021 at 6:32
  • 1
    $\begingroup$ I received a downvote for this answer. A comment would be nice. This is part of what I do for a living. Spacecraft "know" where to point. They typically do not "know" where they are. $\endgroup$ Dec 29, 2021 at 6:35

Your « how do spacecrafts know ... ?» question begs the prerequisite one: do we need them to know [in order to correct their orbits]?. In other terms, do we need them to autonomously determine (a) how they are oriented and (b) where they are (in a given reference frame, at any given time)?

It is quite true (and somewhat intuitive) that we do not wish to, for many missions. As far as we can, we design the spacecrafts to regularly contact the ground navigation center to provide it’s attitude. This is obtained through the spacecraft’s embarked sensors (mainly optical sensors and a stored catalog of remarkable objects called « guiding stars »). By another dedicated radio contact (or the same radio contact, sometimes), the ground team can measure the range and the speed of the spacecraft. With this information the ground navigation center can determine the position, speed and orientation of the spacecraft, which then allow the ground center to program the commands for orbital changes, and/or station-keeping, and/or execution of scientific tasks (which thruster(s) to fire, where to turn to etc...). As a specific example of this approach, this is how the JWST will work. To probe further, read this.

But conversely, there are also many missions, or part of missions, for which the above « earth-centric » approach doesn’t work, or is not optimal. For these missions, we need autonomous navigation, which means the engineers must be able to answer your question: how can we design spacecrafts that can determine autonomously their posistions and speeds (and of course their orientations too)?

Typically, these are missions for which the spacecrafts are too far from Earth. The DART (Double Asteroid Redirection Test) mission is one example. In the final part of its mission, the DART spacecraft must rely on its own capability to determine where it is, relatively to its target, to (autonomously) adjust it’s final course for the most accurate impact. But don’t think that this recently launched DART is a rare example of autonomous navigation. The Deep Space mission is another one. In fact, any mission to Mars must rely on autonomous navigation during the final part of the trip. This the case for the Viking orbiter, back in the 70s …

NASA is also testing the use of reliable cosmic sources as « deep space GPS » for future autonomous navigations (far away).

In fact, long before GPS and long before mobile Internet, our ancestors have learned how to navigate "autonomously" at sea, first by observing the coast lines, then the stars, then using the compass and the sextant. As long as you have good landmarks (in the general sense of visual cues) and a good watch, autonomous navigation should not be a problem (or should be a solved problem with only evolving technical constraints and technologies).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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