Crazycanuk:
A former RAF navigator, now an FTO ground instructor and (when I can afford it) occasional pilot.
Dan Winterland:
I don't want to sound like a smartass, but you did say that you were willing to be corrected. I'll give you Miss Whiplash's phone number later. In the meantime, a few minor points:
The way it finds True North during alignment is just as I said in my previous post. It really is as simple as that - in principle at least. There are one or two practical points in the engineering, obviously, but basically, the idea is so simple, it's unbelievable. I just don't know why most textbooks make it sound so complicated.
Some non-rotating classical INS are cleared for polar flight. It depends on how the software is programmed. If the meridian convergency is dealt with digitally, you can fly over the pole.
All IRSs have accelerometers, to the best of my knowledge. All you can get out of a laser gyro is measurement of angular roatation. With 3 of them, you derive angular rotation from the initial datum in 3 axes. However, this is not acceleration. You still need accelerometers.
As for your final paragraph, not a correction, but an additional tit-bit of information. Astro-trackers were also used on terrestrial navigation systems. The B-58 Hustler's gyro heading system was aligned to a star. The photo-cell was so sensitive that it could actually continue to track the star in daylight, when it was well below the visibility threshold of the naked eye. The original idea was that the system should take a continuous astro position from the heavenly body ( Dammit, there's Miss Whiplash again) but this didn't work very well. However, using the heavenly body as a heading source did prove to be good and enhanced the accuracy available from the gyros of the day.