UPDATEgreenspun.com : LUSENET : CU Statistical Orbit Determination : One Thread
The major pieces of the code are now together and I think that we are finally through the syntax, compiler, column vs row vector, + many other errors in all the routines. The program doesn't appear to want to crash anymore-- keep your fingers crossed, but don't hold your breath. :) THANKS to everyone who responded quickly to my e-mails to get all this fixed. (My e-mail activity increased 100-fold in this past week.) So, we are just about ready for testing to see if the code will spit out decent values.
A few people have gone through their routines to speed things up, but the code is still running slow (and seems to get slower as time goes on). If you haven't gone through your routines to see if you can speed it up, PLEASE DO SO ASAP!! Moriba will have to fine-tune the values in the process noise matrix, and this will be an ugly task if this code keeps running so slow. Not to mention trying to debug any errors in the algorithms. CHECK TO MAKE SURE ALL ARRAYS ARE INITIALIZED TO THEIR MAXIMUM SIZES, UNNECESSARY LOOPS ARE ELIMINATED, AND PERSISTENT VARIABLES ARE USED WHEREVER POSSIBLE.
-- Patrick Shriver (Patrick.Shriver@colorado.edu), April 26, 2000