Hi,
I am looking for this article: Remondi ,B. (1984), Using the Global Positioning System (GPS) Phase Observable for Relative Geodesy: Modeling, Processing, and Results, Ph. D. dissertation, Centre for Space Research, The University of Texas, Austin. Was wondering where I can buy it, or get it from ?
I have been playing with LAMBDA method to fix ambiguities but this method to be reliable requires ADOP to be 10 satellites which gives ADOP bellow 0.5 instantly (I think it is supposed to) but I use gps only chip, so I can not give many degrees of freedom my my equations the ADOP wants.
So there is another method to find correct position which can be used to find correct ambiguities - the Ambiguity Function Method. I have been playing with this one and indeed it is fast and reliable method: I created a cube 30cmx30cmx30cm around true position filled in with points at 1cm interval and I got instantly max of AFM at that position but there is a problem: the bigger the cube the more time the search requires, this small one takes about 1s, 1mx1mx1m takes two minutes. I split the search in four threads now it is 1 minute but I can not get it any better, 8 threads are even slower. One epoch of data takes one minute to process at 1mx1mx1m size. My plan is to use float solutions based on accumulated normal equations for several seconds to bring the position whithin say 3m and than I want to use AFM to search a cube 3mx3mx3m but it takes forever so impractical. The article above explains how to speed up the procedure. I heard about evolutionary search algorithm but this is so much complicated, I need some sort of practical guide...
Any suggestions ?
Thank you
This is a slice of 30cmx30cmx30cm cube, this plane runs through the true position and because of that ambiguity function values are close to1. there are many maximums in the cube (they are like apple shaped if you examine slice by slice in 3D space) but the highest number are only where the true position is. 1mx1mx1m has more maximums than this small one, more apples in the bigger box 😉
I believe you are ignoring too much of the available newer and better technology.
This is an 18 page document by Remondi from 1984:
https://www.ngs.noaa.gov/PUBS_LIB/GPSCarrierPhase.pdf
The original is cited in multiple newer texts, but the following appears to be a source for the original:
This may be your nearest library:
https://opac.unibw.de/InfoGuideClient.ubysis/start.do?Query=10%3d%22BV021918722%22
Paul in PA
Yuly posts over on another forum, that I frequent.
Most of his posts, are over my head.
I'm an end user..... And happy to buy pre made solutions.
I'd never get too far, if i had to make it, before i used it!
Some folks, it's just their calling, to dig in deeper, and zone off into the unknown, of how gps works.
Another way to say it, is if I were in court, and i had to explain the equasions, i use every time i fire my gps system, it'd be a lost cause.
However, some folks are set on such things...
Hey, do it, and make it run faster, deeper, and harder.....
So, my resopnse is a bit like some of my clients.... "You are pretty smart..."
It's easier to buy canned peaches, than to learn to can them....
As Paul in Pa indicates, this is a venerable source.
When I first started studying GPS this was a resource I acquired (in print as bound photocopies of the dissertation) from the NGS Information Center. It is no longer in their list though trying to call them might find one buried somewhere. My copy was passed on to a colleague a decade ago. The Info Center did take many printed documents and scanned them; this one did not evidently make the cut.
BTW, Dr Remondi taught two courses, Intro and Intermediate GPS to NGS and some other federal employees in 1994-96. He left the NGS in 1994. He was an inspiring teacher.
I unfortunately do not recollect his treatment of the issue of integer fixing. I do know that the software NGS used until the late 1990's (OMNI) was semi-automated when it came to integer fixing. A table, like yours, was shown of potential integer unknowns from which the analyst would select. Extensive plotting and manual cycle-slip fixing were also part of the process. Ah, the good old days????
I was surprised that the CSR does not provide digital versions of dissertations earned there.
I will attempt to search my backups and old computers to see if I have a copy of an incomplete version of the dissertation which I believe was subsequently offered by the NGS Information Center.
FWIW,
I copied a PDF version of the article referenced in my earlier post here: http://geodesyattamucc.pbworks.com/w/page/37089695/Links#view=edit
Cheers,
GeeOddMike, Paul in PA and Nate thank you very much ! for contributing and time.
it started from here (not sure if link works), print screen bellow:
Another researcher Azmi Hassan did improve the search, this one sounds really fascinating: ambiguity function values (see my pic above) are treated like a population of individuals (also on all below and above planes too), and their relationship in some vicinity is used to find all maximum areas:
But I still cannot find any details how both researches exactly, in great details do it, so this question is still open (but I have not finished reading
the articles you linked to, may be it is there 🙂 Thanks.
GeeOddMike it is amazing that you know Benjamin W. Remondi personally..
Yuriy
Evolutionary algorithm should solve the problem (now I need c++ implementation of EA) this link describes the algorithms:
http://www.solver.com/genetic-evolutionary-introduction
This is from the description: The use of a population of solutions helps the evolutionary algorithm avoid becoming "trapped" at a local optimum, when an even better optimum may be found outside the vicinity of the current solution. this is exactly my case.
It is interesting that this is not a GNSS problem anymore is a more general problem.
Amazingly it did not take long to find an implementation: http://eodev.sourceforge.net/
This is indeed fascinating !
