Have you compared the time & accuracy results from tweaking your rinex file as against re-observing the GPS points?
In my opinion GeeOddMike pretty well nailed it in his first post. Back in the day it was worthwhile tweaking data by excluding satellites, portions of "bad" data, playing with the horizon mask, deleting entire time segments etc. however software is immeasureaby better today. All of it pretty much takes care of all that for you, so it's scarcely worth the time anymore. I would advise throwing all the data in there even if some of it does have have GDOP values and let the software sort it out for you. Frankly it will probably do a better job than you or anyone else can.
To me, the secret of successful GNSS baseline processing is to get enough data to begin with. Stay on the point as long or longer than you think you need to. If you have enough time on the point the software can pull out a good solution. If you don't have enough time on the point, you or the software can play with it endlessly and you still won't get a strong solution. As a trainer for Leica Heerbrugg told me years ago when discussing some of our data, "You can't make gold out of @#$%!"
All the software out there seems to work quite well today, but I will say that, at least a few years back, the best by far was GrafNav that I used embedded within the Aplanix software for lidar processing. It was pretty amazing in its ability to generate a correctly fixed solution and hold on to it. We could fly up to 150 km. from the base at the airport and sometimes still be fixed. When we lost fix, it was uncanny in its ability to successfully re-establish the correct integer ambiguities even tho the aircraft may be a long ways from the base station. Most software is probably that good these days anyway. A huge improvement happened when Peter Teunissen at Delft University developed the LAMBDA method for fixing ambiguities. Now everyone uses it and getting a fixed solution is almost trivial. A few years back Yehuda Bock even had kinematic software that would successfully fix ambiguities to integers independentally for EVERY one second epoch. Anyway, double difference processing has been around for quite a while now and the software has evolved tremendously from the early days when software would simply round off the real number floating ambiguities to integer values and would even let you have a guess at fixing them yourself. None of that worked, by the way.
A few years back Yehuda Bock even had kinematic software that would successfully fix ambiguities to integers independentally for EVERY one second epoch.
I was surprised by the reference to "a few years back," as I assumed that the Epoch-By-Epoch product was still available. I just
looked at the Geodetics Inc. website, and they've gone over to producing hardware, with no software listed. Yehuda's name isn't even listed, it looks like his wife is in charge.
BTW, Yehuda Bock's JGR article is available for free viewing/download at:
http://onlinelibrary.wiley.com/doi/10.1029/2000JB900268/epdf
I remember being quite impressed by the idea. With a solution for each epoch one could use standard statistical analysis to derive the best coordinates rather than relying on integration over time spans. Unfortunately, I never followed its development and use as a product.
I imagine the "women owned" status provides some benefit for government contracting.
Hope this contributes,
DMM
Thanks for the listing of Bock's paper about every epoch ambiguity resolution. I re-read it and found it interesting that Teunissen actually presented his LAMBDA method as far back as 1993. Geodetics was, and I guess still is, an amazing company. They used to offer a real-time network software that used their every epoch initialization approach. We seriously considered purchasing it when I was with the City of Albuquerque a few years back. They were in town for a convention and I talked to them quite a bit about all this. They were certainly ahead of the pack in research and had some amazing methods for analyzing data and estimating tropospheric and ionospheric delays. According to the article their every second cleaned data gave six hour solutions that were as accurate as twenty-four hours of batch processed data. Pretty amazing. I was surprised that no one else got on to the idea of every epoch resolution; however, it seems that these days ambiguity resolution has become almost trivial and there simply isn't enough advantage to make every epoch resolution worthwhile. Interesting stuff, and, by the way, Yehuda Bock died a few years back.
I am surprised to learn of his death. I note that the Scripps site does not indicate it. See: http://scrippsscholars.ucsd.edu/ybock/biocv
As I retired in 2008 (fully in 2013), I am officially out of the loop, pursuing my interests much less systematically.
As far as the epoch-by-epoch approach, I had thought it might be especially relevant for heighting given the known deficiencies in tropo modeling. I was involved in a project attempting to link GPS and tide and water level sensors. We saw the applicability of his approach to our problem. Unfortunately, funding and logistic problems continued up to my departure.
Cheers,
DMM
My first post here, after a fair amount of lurking ...
Dr. Yehuda Bock is very much alive and actively working in geodesy at Scripps Institution of Oceanography, in CA.
Here's a link to a Nov. 2 Washington Post article referencing him:
https://www.washingtonpost.com/national/health-science/when-an-earthquake-is-coming-how-can-you-get-even-a-little-warning/2015/10/30/df12c640-63c1-11e5-9757-e49273f05f65_story.html
Dr. Bock might well echo Mark Twain's sentiment, along the lines of: "The reports of my death have been greatly exaggerated." 😉
Bill
I was pretty surprised when I heard he passed. I got an Ntrip account through him just about a year ago.
Glad to hear Dr Bock is still alive and kicking. I commend his great contributions to geodesy and lesser fields.
Thanks for the correction and update Mr Stone. Welcome to the board.
Potentially offensive Python link follows (Monty that is):
[MEDIA=youtube]Jdf5EXo6I68[/MEDIA]
Hey Bill
Are you the William Stone I did the OPUS project with along with Mark in Sisters?
John Minor
Well, I hope Dr. Bock did not hear about his early demise. I heard about it some time back and certainly got something wrong.
RCliffWilkie, post: 343627, member: 10285 wrote: Well, I hope Dr. Bock did not hear about his early demise. I heard about it some time back and certainly got something wrong.
He was quite ill for awhile, so that may account for the rumors.
Well, I'm glad to hear Yehuda Bock is alive and well. Hope he has many more years to come. I heard of his alleged demise about eight years ago and had no reason to follow up with any more information about it. I never met him. I met his wife Lydia and several of the employees of their company Geodetics about eight years ago when they were in town for a convention. We met so they could tell me about their real time network offerings. Not long after, a friend, who did know Yeduda Bock, said he had died from cancer. Or at least that is what I thought he said. I had no reason to doubt it or to investigate further. I do vaguely recall reading several papers by him. The one cited previously about every epoch initialization and several, I believe, about mm. level deformation modeling using GPS. They were paralleling some work done by Ben Remondi doing the same thing using triple differencing techniques. It did seem odd that Yehuda's name was not mentioned anymore on the website of Geodetics. Regardless, here's to a long and happy life for Yehuda Bock. Hope to see more research papers by him soon. He certainly had a lot to offer.
