I’ve been reading up on this but don’t really get it.
The stars in the sky look the same regardless of where the earth is in space – because the stars are far away in relation to the diameter of the earth’s orbit around the sun, and the sun moves through the Milky Way etc very slowly.
So taking angular measurements on multiple stars only gets you information on which way the aircraft (or spacecraft) is pointing. The angle between any two stars will always be the same no matter where in space you are (the parallax of a star can be measured over a year or so but is tiny). It doesn’t tell you where you are on the earth in 2D or in space in 3D.
I can see that with a “DME” added, it works, but for autonomous navigation?
If you mean celestial navigation it is a form of triangulation. You measure the height of certain celestial bodies versus the horizon and it creates a few lines which should intersect at your position.
Eg. If a star is over a point on earth and you see it as 35 degrees above the horizon, that implies a circle of position.
But you need an accurate clock as the tables are all driven by time.
The SR71 had a pod built into the upper body for astronavigation. Pretty interesting stuff.
I can see how it works over the earth’s surface, with an accurate clock.
At its simplest, with a sextant (an angle measurement tool), you wait till the sun is highest at mid-day and you measure its angle and that gives you your latitude. Then, if you have accurate time, you can calculate your longitude (not sure how exactly that bit works).
And you can do the same with a star which is somewhere in the north or the south.
What I don’t understand is how the SR71 system worked. Lockheed were using that on spy drones in the 1960s, which like the SR71 flew high enough to see the stars anytime. That system just looks upwards, at a single patch in the sky, and had a pattern matching feature which AFAICS would enable a very accurate heading to be flown. But not an accurate track.
It’s not the position of the earth in space which matters (which indeed requires angular precision beyond what’s achievable on a ship) but the orientation of the earth towards the sky, the configuration of which is known for a given time and place (hence the need for a good chronometer). So for a given time, there will be only one Great Circle from which a given star will be at a certain elevation to the horizon). If your watch is off by a minute, you’ll be off by 15nm!
With just the sun you can get your latitude pretty easily, by measuring its angle above the horizon when the shadow is shortest or aligned north-south if you have a compass; but you need to correct for the date, because the inclination of the earth means the sun is right above the equator only twice a year. With the sun and a chronometer set to GMT, you can measure the time difference between you and GMT (shortest shadow is midday for you), which is your longitude.
The degrees a star or planet (extra tables) does come up over the horizon at a given local time/date gives you a closed line for your position (you need a table that corrects for earth orientation), if you know latitude (compass/polar star) or longitude (universal UTC watch) you can get an exact position fix
If you use sun/moon (which tend to be opposite) you will know your position from the two “position lines” as long as you know which half of earth you sit on: if you are really that unsure you can use polar stars by night, winds or just check for penguins vs auks, the latter flies in water and in air, so if you see one then you are not far from the north pole ;)
A third way to say what denopa and ibra say, and applied to “pattern matching” stellar navigation:
- if you know the pattern of stars you are looking at, you know exactly which direction in space you are looking.
- if you know you are standing on a (roughly spherical) non-spinning body looking “up”, that gives you exactly one point on that body you can be
- if you know the shape, orientation and spin parameters of a spinning body and the time, that will give you exactly one point on that body.
So in summary, stellar navigation uses time and establishing where your Zenith is on the celestial sphere to pinpoint your position on Earth.
Curiously enough, without looking at the orientation of the stellar pattern, this does not give you a heading (although you get that “for free” when pattern matching)
if you know the pattern of stars you are looking at, you know exactly which direction in space you are looking
On the earth, or not far above it, that gives you a direction in space, as you say. This is like flying a heading with a compass. Then how do you correct for wind drift? You can drift sideways and the star image will look the same, because the stars are “at infinity”.
if you know the shape, orientation and spin parameters of a spinning body and the time, that will give you exactly one point on that body
I am not so sure. You would have to be extremely accurately perpendicular to the star image for that to be the case; I guess the star nav systems try to achieve that. You would need to get the star light to shine through a template which is some distance away from the sensors; basically it would have to pass through two templates, to ensure this.
Looking at a star alone does give you the direction in space of the star. You can only use that to determine a heading if you know the time, with the only exception being a visible star exactly in the celestial north or south pole. Polaris is accurate to 0.5 degrees so you only need the time if you want a more precise heading.
If you look up vertically, it gives you a vector in space that immediately translates to a point on earth, using time. You are absolutely correct that for this to work, you have to be pretty accurate in knowing where “up” is.
If you don’t look up vertically, you can of course use the elevation (angle above the horizon) of the stars you see and the time to determine where the Zenith is.
If sufficiently interested the RYA ocean yacht master or equivalent would be ideal for you. Id expect at sea and not in practice accuracy to within 10 miles, reducing down to 5 or better. There are those who consistently achieve a mile or better, although the less the deck is rolling ….. There is no substitute for a good sextant. The trouble is at sea as in the air we all resort to GPS, but its good to know the alternative will get you across an ocean with sufficient accuracy not to miss a spec of an island.
