[AR] Re: Star Trackers, was Re: Re: Spin stabilized rocket

  • From: Henry Spencer <hspencer@xxxxxxxxxxxxx>
  • To: Arocket List <arocket@xxxxxxxxxxxxx>
  • Date: Tue, 25 Dec 2018 15:42:59 -0500 (EST)

On Tue, 25 Dec 2018, Monroe L. King Jr. wrote:

My first question is can you test a Star Tracker from earth?

Yes.  We've tested star trackers outside on a clear night.

Can it get a position from a single frame?

Yes. The Sinclair star trackers, in particular, solve the "lost in space" problem -- no idea what the attitude is, figure it out from the image -- every time. Some other designs do that the first time, and then do faster or more precise updates using a different algorithm that starts from a known approximate position.

I also wonder if the spacecraft is tumbling can it get a position? What
roll rate and tumble can a star tracker handle?

That's part of the star-tracker specs. They all have an attitude-rate limit -- as do most other attitude-determination methods -- it's just a question of how high it is. Spacecraft can generally live with fairly low limits, because they don't often get into *fast* spins (and if they do, generally they can use some less-precise sensor, like a MEMS gyro, to first bring the spin rate down to something the star tracker can handle).

The camera would matter in this right?

For sure. One important question is, how long an exposure does it need to get clear star images in the worst case? (There are some areas of the sky, especially the southern-hemisphere sky, that don't have much in the way of bright stars.) This tends to set the attitude-rate limit -- when the motion is fast enough to turn star images into trails, longer exposures don't help, because they don't make the trails brighter, just longer.

This also means that when designing the camera, more pixels -- which usually means smaller pixels -- is not necessarily better. The CCDs in the Clementine star trackers were 576x384.

...if a star tracker can be used from earth (this is why I'm supposing they can't work from earth) Why don't telescope tracking systems use it?

Never looked at them, but one thing that's likely to be an issue is that solving the "lost in space" problem quickly, with a star catalog of reasonable size, typically requires a rather wider field of view than the usual astronomical telescope.

(Another noteworthy issue, by the way, is that errors in star catalogs are not uncommon. Yes, including the Hubble Guide Star Catalog.)

...What kind of math does a star tracker use?

Depends somewhat on what algorithm it uses; there's more than one answer.

My last question at the moment is: how difficult is it really to go from an earth frame of reference to an inertial, body fixed from an earth fixed frame?

Actually calculating coordinate transformations is pretty straightforward, although getting a full understanding of the details -- so you know what to compute -- can certainly be a headache.

Henry

Other related posts:

  1. Home
  2. arocket
  3. Archive
  4. 12-2018