Date: Wed, 8 Oct 1997 11:40:53 -0700 (PDT) From: Dave Babler <dbabler@Rigel.orionsys.com> To: Nate Williams <nate@mt.sri.com> Cc: Mike Smith <mike@smith.net.au>, cliff ainsworth III <cliff@cliffsworld.com>, freebsd-hardware@FreeBSD.ORG Subject: Re: project truck.....ideas wanted Message-ID: <Pine.BSF.3.96.971008110544.1345B-100000@Rigel.orionsys.com> In-Reply-To: <199710081658.KAA10961@rocky.mt.sri.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 8 Oct 1997, Nate Williams wrote: > > > > Damn, I forgot DGPS. You will want a decent wireless radio link > > > > > > You don't even *need* a wireless radio link, since the GPS unit will > > > allow you to synchronize your clocks with the satellites, thus allowing > > > you the ability to use time-stamps for your readings that you can > > > 'differentialize' after the run. > > > > Uh, hang on a second. You want to use DGPS to remove the SA jitter, > > Shh, don't tell anyone, but let me let you in on a little secret. The > 'jitter' in GPS that makes most of the difference *isn't* random. Most > of it is introduced, and that's what makes the accuracy < 100 M. If you > remove the accuracy, you're down to ~1M accuracy, and by taking out the > rest of the 'jitter', you can do better than .1M accuracy. However, for > the above application, I suspect 1M accuracy is probably good enough. > > We call this 'fake' DGPS, and use it for many projects at SRI which > don't need *really* accurate measurements. (SRI helped develop GPS, and > continues to do alot of GPS research.) > > Again, I'm talking about removing the 'introduced' jitter, and not > removing the jitter related to moving satellites/moving cars, but I > don't think that kind of accuracy is *necessary*. > Unfortunately, you really need to say "maximum possible accuracy" using any form of post DGPS correction, especially for live real-time measurements. You can also often use "real" DGPS by adding a DCI correction unit and subscribing to their correction service, though you are limited by their coverage ranges (they use commercial FM radio stations to broadcast the correction factors). SA is easy to remove, but it is not what ultimately determines accuracy - other considerations do, such as satellite azimuths, multipath distortion, atmoshperic distortions and so on. Post-measurement DGPS is less accurate the further apart the measurement GPS and base GPS are, since they will possibly be seeing a different constellation of satellites at least some of the time. One advantage of DCI correction is that it includes corrections for all of the visible satellites, not just the 4 that happen to be used for the current 3D position fix. In developing a GPS-based position-reporting mobile data gathering system using both "true" DGPS and base-station DGPS we took a fair amount of field measurements. For "live" data - one sample every 5 seconds in our case - by far the largest errors occur in fixed base station post correction when the constellation of satellites change - often introducing a sudden 10m apparent change in position for a few seconds as the new satellite factors cranked in and the position came back to within a meter or so of the "true" position. These interrim readings *are* marked as less accurate by the GPS receiver, but if you must have a stream of live measurements you may not be able to just wait the minute or so for the accuracy to return. For vehicle readings, placement of the GPS antenna can be important, especially if you are using other radio antennas for telemetry. All of our experience was using Trimble Navigation 6-channel OEM units, but the specs are fairly close to units from Rockwell. -Dave Babler
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.96.971008110544.1345B-100000>