Yes that is still radio nav based but you will need to have a bunch of large pulsars detectors and huge compute power onboard,
At the moment what we sent beyond earth/moon are small probes with GA aircraft size and 1KW max power, so they will still rely on ground-based navigation than having something onboard? But not sure if you can still use that for onboard position fix and send data (1kb/s) to earth for trajectory calculations?
Someone sent me this file_pdf which suggests you could use x-ray pulsars. It is heavy going maths-wise but there are many interesting snippets e.g. on page 7 that with appropriate corrections the timing can be as accurate as an atomic clock.
The report also states that the present Deep Space Network can determine a spacecraft position within 10km at a distace of 1 AU (the distance from the earth to the sun, roughly) which is pretty amazing.
“how do you navigate in deep space?”
Most of the times you don’t, you just have to trust free fall equations as by the time you reach deep space there is nothing left in power and things will go exactly as expected,
For trajectory tracking, we still rely on radio navigation with signal from the spaceship, these are received by earth ground stations (2 or 3 stations) then all gets processed on ground rather than onboard, while DME can handle max 100? aircrafts before saturation, I think “NASA DME” does 1 aircraft at a time ;)
My next Q is: how do you navigate in deep space?
I know with Apollo they used earth based radar to confirm the spacecraft position. Amazingly this was usable even close to the moon. Obviously this won’t work much further out.
Many years ago someone involved in the satellite launching business told me they use DME distance, and knowing the orbital equations this is all they need to work out where it is. They get thousands of readings and end up with thousands of simultaneous equations which are solved and the 3D position pops out.
But how could a spacecraft do this autonomously?
Peter wrote:
They did have a reasonably accurate clock on board, though only quartz crystal controlled, so probably within a few seconds.
For the length of an Apollo mission (only a few days), a quartz crystal controlled clock should remain within a second, especially with a TCXO.
…I beg to differ… 
It does exactly what all stellar navigation systems do – establish a definitive vector in space. It just uses pattern matching instead of fiddling with a sextant to achieve this.
This can be used to determine the orientation in space of the instrument, and on Earth, together with time, the location.
There have been three types of answers in this thread
Elon Musk’s Starlink system (space-based Internet communication) uses a star tracker for precision pointing. It has only little in common with celestial navigation using a sextant.
In order to get anywhere with celestial navigation you need five things:
1. A device to accurately measure angles above the horizon
2. A clock giving you an absolute (current) time reading
3. Lookup tables or formulae to convert your angular and time readings into a Lat/Long position
4. A visible horizon
5. A visible astro body fitting your tables/formulae
At long sea voyages you usually have (4) and (5) often enough (unless stuck in fog, haze or under clouds for extended periods).
(3) you can prepare ahead of time before even taking off.
(2) used to be a tough problem but can be considered „solved“ for most practical purposes by investing 10 bucks in a quartz watch.
That leaves (1) as the only „real“ problem, which is typically taken on by relying on a sextant.
Unfortunately a full-fledged sextant can be quite a tricky beast to handle properly. Not to speak of the cost, weight and sensitivity to all kinds of environmental factors like temperature, humidity or shock.
One way to get around all these issues around classic sextants is something like the „Bris Sextant“ (https://en.m.wikipedia.org/wiki/Bris_sextant), which has no moving parts and is small enough to be worn on a necklace!
The basic idea is to give up on the goal of trying to accurately measure any angle, but rather only a certain set of convenient angles.
For example, rather than measuring the angle between the sun and the horizon right now and then recording 26.7 degrees and the current time one can also simply wait until the angle is exactly 30 degrees and record the time at that particular instance.
In many cases this trading the ability to establish a position at any time with complicated and costly machinery against doing the same only at particular times but with a great simplification in measuring and equipment can be quite beneficial.
Especially since „anytime“ isn’t „anytime“ anyway, since you always need a horizon and a clear sky, even with the best and most expensive sextant in the world…
And: By reducing your possible angular readings to, say, only 8 or 16 cases your lookup tables become a lot smaller, which can be quite an important benefit as well.
Unfortunately, waiting for the next convenient angular reading time is much less practical in a fast moving aircraft than a slow moving ocean vessel.
In the plane we simply can’t wait two hours for the next „positioning event”.
But for sea or even space navigation I find this basic principle quite interesting…
Was Amerigo Vespucci the first to discover where America was, using moon and star observations?
Other astronomers with his skills stayed at home at that time.
To use Astro Navigation, you calculate the position of a body so as to determine exactly where it will be overhead on the earths surface at the desired time. If the body is exactly overhead the Elevation will be 90 degrees. As this is rather limiting, you calculate what the Elevation will be at a given position know as an Assumed Position. If the Elevation you measure is greater or less than that calculated, you will be displaced along the bodies Azimuth by 1 nm per minute of arc from the assumed position.
You do not measure the Azimuth, you calculate it to locate the body. You only ever measure Elevation or “Altitude” using the sextant. One observation gets you a single position line at 90 degrees to the Azimuth. 2 bodies will produce a sandwhich fix with lines at right angles whilst 3 bodies having Azimuths displaced by 60 degrees will produce a cocked hat. You can also take sightings of the ground using a sextant to determine range.
Because of aircraft and operator movement, a number of shots are taken and averaged over one or two minutes. Sometimes 3 or 4 seperate shots will be taken on the same body and the average taken (V Force technique)
