NavList:
A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
Re: Unistar Principle etc.
From: Alexandre Eremenko
Date: 2012 Apr 21, 00:56 -0400
From: Alexandre Eremenko
Date: 2012 Apr 21, 00:56 -0400
Dear Kermit, > Please do explain the Unistar principle to us. I'll try. I cannot say that I recovered all mathematical details. (And even less, the engineering details). As I said, there is no satisfactory exlanation in the literature. > And I bet that your will find that the star which > is to bring the highest accuracy update is ... just the one on top of > the target when the missile reaches it. I afraid your bet is wrong:-) The adjustment from the sight is made just before the engine switches off. That is about 6 min after launch. So it is the position of the star at the time of launch, at the place of launch which is relevant. They point the "sextant" at the star before the launch. The star you propose will be too low, (and even can be obscured by the the Earth, if the target is 6000 miles away at this moment:-) According to one account, the best star is in front of the of the missile at the moment when the engin is stopped. My "theory" shows that any star which is reasonably high over the horizon and has approximately the same azimuth as the target will work. (Will work from the mathematical point of view. There are also engineering restrictions on the size, placement and width of the window in the missile). But let me try to explain what I understand. There are 9 parameters whose adjustment can be relevant: 3 coordinates of the missile (say, latitude, longitude and altitude). 3 components of velocity. 3 parameters of what they call "attitude" which characterize the orientation of the missile with respect to the stars (more precisely, the orientation of the gyroscopic platform). The first 6 parameters determine whether the missle will hit the target or not, and their value at the moment when the engine switches off is important. (The engine works for about 6 minutes, after which you cannot change anything). Shortly before this moment, the star is sighted, and the last 3 parameters (the attitude) are only needed to recompute the trajectory and adjust the first six. The observation of one star gives you two numbers (for example, altitude and azimuth). If you measure altitude and azimuth of one star exactly, with respect to the frame connected to the Earth, then you derive your position exactly. The problem is that you measure not with respect to the Earth related frame but with respect to your gyroscopic platform frame, which can be off by some error. The problem is how you adjust many (9) parameters from only 2 measurements. One account says that they do this statisticaly. The relative importance of these parameters is not the same. They collect the empirical data from the tests and make on the base of these data a 9 times 2 "correlation matrix" which gives adjustment of 9 parameters from 2 measurements. This looks like a very naive and non-efficient way to solve the problem, and I do not believe that this is really the way it works. Here is a simple argument which shows that information in two parameters from the star sight should be sufficient for exact adjustment. Suppose you do not adjust anything. Then you miss, your missile hits the ground somewhere near the target. Where exactly it hits is described by TWO numbers: miss in distance and miss in direction. That's why two adjustments must be enough:-) According to some authors, it is only the direction of the velocity vector that is adjusted. The optimal star for this is in front of the missile. When I look at this more closely, I find the following. First of all, 2 of the 9 parameters cannot be corrected from any number of star sights, namely the altitude of the missile over the sea level, and its speed (speed is the magnitude of the velocity vector, it is just a number, mesasured in miles/sec). Because from the two missiles flying at different altitudes at different speeds, the sky looks EXACTLY the same. Of course, altitude does not really have to be adjusted from the star sight, because there are other good ways to measure it. Speed is a different matter, but in no way it can be measured from a star sight. (They make ONLY ONE sight, there is no time for two, and no space to set two sighting devices). I could solve comletely the simplified problem where everything is in one (vertical) plane: the center of the Earth, the missile, the target and the star. In this case, second star does not help, but EXACT adjustment cannot be made. I suppose the same is true in the general situation. Where is the star in this plane does not matter (privided it is high enough, so that refraction has no influence. Complete justification of this needs some formulas which I am reluctant to write:-) One account says that the pure mathematical problem caused some difficulty to the designer team... But much more difficult are the engineering problems, and one of them is how to make the window through which the sight is made:-) It must be strong enough to stand the launch (the active phase lasts about 6 minutes), but perfect optitally, not to distort the sight. Another problem is how to make sure that the telescope finds the right star. They also pre-compute all possible occultation, to make sure that the selected star is not occulted by the Moon or Sun or some planet. The telescope is supposed to find the star in the daylight as well as at night:-) It helps that the missile is already high enough at the time of the sight, so the sky is not blue but dark. This message is too long, but I can continue... Alex.