Article: 143469 of sci.geo.satellite-nav Sender: wolfgang@capsicum.wsrcc.com Newsgroups: sci.geo.satellite-nav Subject: Re: DGPSIP - What now? References: <8euoq6$tqo$1@mail1.wg.waii.com> From: Wolfgang Rupprecht Message-ID: Organization: W S Rupprecht Computer Consulting, Fremont CA Lines: 104 User-Agent: Gnus/5.08 (Gnus v5.8.5) Emacs/20.5 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Fri, 05 May 2000 19:51:52 GMT NNTP-Posting-Host: 24.1.65.208 X-Complaints-To: abuse@home.net X-Trace: news1.frmt1.sfba.home.com 957556312 24.1.65.208 (Fri, 05 May 2000 12:51:52 PDT) NNTP-Posting-Date: Fri, 05 May 2000 12:51:52 PDT Path: news1.meer.net!news3.best.com!news1.best.com!newsfeed.mathworks.com!newsfeed.direct.ca!newshub2.rdc1.sfba.home.com!newshub1.home.com!news.home.com!news1.frmt1.sfba.home.com.POSTED!not-for-mail Xref: news1.meer.net sci.geo.satellite-nav:143469 Marc Brett writes: > DGPS over TCP/IP (DGPSIP) seemed to give visible improvements over even > absurdly long (say transatlantic) baselines. > > Now that SA is no longer the dominant error, will very-long baseline DGPS > still improve measurements, or will it even degrade them, since the > base station and the rover are looking at the same satellite through > very different parts of the sky? I'm seeing a degradation of 2.5 meters with a 35mile baseline. %-) If one ignores the above offset, I suspect that the useful radius of dgpsip dropped approximately in the same ratio that the "noise circle" of the gps dropped by. Perhaps there are some tricks that I can pull by getting a WAAS receiver and gatewaying that into the internet instead of the local differential signals. That would allow me to tailor the dgpsip to a geographic area. But that assumes that WAAS still does anything useful at reducing gps positional noise. Anyone with a WAAS setup care to comment? DGPSIP - A Look Back DGPSIP had its first birthday on Jan 17, 2000. Thats a year after the first server sent TCP/IP RTCM corrections from a Garmin GBR-21 across 10 feet of ethernet to another server (running in client mode) which sent the corrections onto a Garmin 12xl. On May 1, 2000 the government turned off selective availability (SA). That basically allowed any GPS to have roughly the same accuracy as a differential gps that was listening to one of the free coast guard DGPS transmitters. Overnight (literally!!!) dgpsip became superfluous. Am I sad that DGPSIP instantly became a dinosaur? Well a little. Its real purpose was political though. I wanted to be able to show the average person how badly SA was screwing up the signal of their GPS. Even long baselines seemed to have much of the SA "wandering" eliminated. It was a great tool for showing how "quiet" things would get when SA was off. The problem was that even "short" 1000 km baseline could have short term biases of 8 meters or so when the sun was up. So in effect it gave someone a tool that allowed them to do quick short term relative measurements and so-so longer term measurements. (At these distances one could get better accuracy by doing a 24hr average of GPS's position.) After the first version of the code was released I started to get lots of very enthusiastic EPE reports from all around the world. I didn't want to quell people's enthusiasm for experimenting -- after all I get a big kick out of turning people onto science and math. I did suspect that folks were waiting for hours to get the best numbers and then dashed those off to me. While it is nice to know how well something could work at its finest, I really wanted to know how well it would work on average. I added the built-in standard deviation calculator to the remote code and had the remote send back the calculated deviation on the lat, lon and height. For folks with a good sky view and within a 3000 km radius the standard deviation numbers were always better by a factor of 3 than the no dgps SA-active case. What was most interesting about the whole experiment was the number of connections I got *.mil and *.gov sites. At first it really weirded me out, especially the ones from bases that were controlling some aspect of GPS segment. ;-) Then I started getting friendly email from folks at the gov/mil sites asking how they could run this or that aspect of the various dgpsip programs. All the folks that sent mail were into using GPS's recreationally and were pleased as punch that their personal consumer-grade gps could be made to work as well as the military gear. The email that made me smile the most came from a mil site that works directly with the gps SA errors. The operator made a point of mentioning that whenever the policy-makers come by the site he shows them the dgpsip server and points out how it works over a very large area. I also got a request from a military training center that wanted to use a private dgpsip server running on their base behind a firewall to distribute correction signals to the fleet. That one made me laugh. I assume the rules for dealing with the crypto were so cumbersome that they would rather deal with RTCM corrections. Well, its certainly been a fun year and a third. I don't plan on turning off the server any time soon. (After all what do I need the differential radio for now anyway?) I would like to thank the 10's of thousands of folks that have gone to the trouble to download the code and use the server. I'd especially like to thank the folks that have set up remote feed sites. These were invaluable for showing folks how good things got when you were really close to a reference site. Lastly I'd like to thank all the folks that took the time to send email and ask interesting questions. The best way to learn something is to try to explain it to someone else. It is amazing how many holes and problems that uncovers with one's knowledge. I've learned a quite a bit about GPS that I otherwise might not have. Thank you folks! Its been fun! -wolfgang -- Wolfgang Rupprecht http://www.wsrcc.com/wolfgang/ DGPS signals via the Internet http://www.wsrcc.com/wolfgang/gps/dgps-ip.html