A GPS question

Take a look at this publication
see table on page 19

Here is another reference

My understanding is that an sv's broadcasted location is more of a function of it's relationship with the other satellites than with predicting where its at at any given time during it's orbit.

In other words, they constantly talk to each other and millions of calcs occur every minute while they broadcast their position 'solutions' to us here on the geoid.

From watching the NANUs I believe from time to time an sv's speed or altitude is adjusted slightly and is usually back 'on-line' within 24 hours.

In other words, they depend more on their distances between each other to calc their own location than anything else. Still quite an accomplishment though.

Howdy,

Not sure I understand your question. As for the question of how accurately we know a satellite's position, see the following table:
http://igscb.jpl.nasa.gov/components/prods.html

Keeping the satellite orbits stable and generating broadcast orbit ephemerides is a DOD function. Precise users of GPS use the a posteriori orbits available from the International GNSS Service. The DOD does generate precise orbit products in the WGS 84 reference frame which differs from ITRF which both differ from NAD 83.

If your question is how much of an improvement in satellite position requires reprocessing with the better orbit product, use the following rule of thumb: baseline accuracy divided by baseline length is proportional to satellite position error divided by the distance to the satellite.

This assumes you are doing positioning with respect to a known site. If you use mostly short baselines, < 10 km, current broadcast orbits are fine. I always process with at least the IGS rapid orbits.

Hope this helps,

DMM

> Keeping the satellite orbits stable and generating broadcast orbit ephemerides is a DOD function. Precise users of GPS use the a posteriori orbits available from the International GNSS Service. The DOD does generate precise orbit products in the WGS 84 reference frame which differs from ITRF which both differ from NAD 83.

Yep, the DOD doesn't move the satellite, they guess (really good) where it will be, then they publish the (a posteriori ephemerides) actual positions they were in.

> Yep, the DOD doesn't move the satellite, they guess (really good) where it will be, then they publish the (a posteriori ephemerides) actual positions they were in.

You're probably thinking of the IGS (International GNSS Service) orbits that are computed after the fact.

From the IGS website :

"The International GNSS Service (IGS), formerly the International GPS Service, is a voluntary federation of more than 200 worldwide agencies that pool resources and permanent GPS & GLONASS station data to generate precise GPS & GLONASS products."

it would be interesting to know how much a satellite is allowed to go off its design orbit before they use its thrusters to put it back, that must be tens of meters, I guess.

Broadcast Calculated Orbits Give Close Raw Positions

With a known fixed base point on the ground one gets very good differential positions.

Once one downloads the ultra rapid after the fact orbits differential positions get precise.

Waiting till the next day, one can download the rapid after the fact orbits differential positions get very precise.

If one waits 2 weeks for the precise orbits it is difficult to see any improvement in precision.

OPUS is simply differential GPS meaned from at least 3 control positions. It improves in the same manner.

In the past I always waited for the rapid orbit to establish my local control sites. In this past year almost all projects had local control sites established from the ultra_18 orbit, available in the early evening after a day of field work. For those that I later checked against the rapid data, few required any change. I have seen some changes in the 5th decimal seconds position, but hey that is less than my setup error allowance.

When I say a fixed base ground point, one can actually read a position and elevation off a quad map and be sufficiently precise for very good differential measurements. Now raw Lat and Lon from GPS receivers is better than quad scaling but the raw elevation is still loose. Using a quad derived elevation one fixes the GPS observations to the face of the earth and very good comparisons can be made with conventional traverse even using the receiver derived orbit data.

Ultra rapid orbits are available during the day of observation, igu*****_00.sp3, _06.sp3. _12.sp3 and _18.sp3. The availability time is about halfway through the time period, 3 hours after start time. So the _18 is nominally available at 21 hours UTC or 4 PM EST. It contains actual orbit data till 18 hours UTC or 1 PM EST plus 6 hours of revised calculated orbits. A calculation from a known position can be very good in the short term. They are working on providing the rapid orbit faster than noon the next day, but it involves some additional cross checking. As far as survey work goes today's rapid orbit is the precise final orbit.

Paul in PA

Think Kilometers Yuriy

Those satellites are 20,000 KM over your head. Even at that height there is some atmospheric drag that slows them down. As they slow their orbit drops. But as they circle the earth they are also bouncing around up and down due to local gravity, the sun and the moon variations.

From the center of the earth the orbits are in the 26,000 Km range. One orbit of the earth is 163,000 Km in circumference. There are six nominal planes for the satellites and at orbit elevation they are 27,000 Km apart, satellite to satellite/plane to plane. So there is a lot of room for them up there. Assume we have a pefect plane with 6 satellites, 27,000m apart. If a satellite is 10,000 Km out of position behind it does not really affect the constellation geometry, but it indicates the satellite is slowing and losing elevation. A bit of boost in speed will raise the orbit but it does not take much. Assume the speed was increased only 5 Km/Hr, it will take 2,000 hours, about 3 months to get where it should be. Do you know where you will be in 3 months?

There are 31 active satellites up there, but that is not all. There are also satellites that could be made active if necessary. The nevest satellites are SVN 62 and 63, the IIF models, with a third scheduled to be launched next week. The ten lowest active SVN numbers are 23, 26, 27, 33-36, 38, 49 & 41. They were not always launched according to their build number. I surmise their are at least 6 such spare vehicles up there, meaning at least 6 SVs in each orbital plane. SVN = Space Vehicle Number different from the PRN number used in data broadcast.

Paul in PA

The question of when the satellite orbits are adjusted is an interesting one.

A satellite in orbit has two forces keeping it there. It has a velocity perpendicular to the earth pushing it away from the earth and the pull of gravity from the earth causing it to fall towards the ground. If these two forces are equaled out then the satellite obtains a stable path or orbit circling the earth.

However, that’s just the beginning simple version. The earth is hardly a perfect sphere and there are different gravitational forces acting on the satellite depending where over the earth it is. More importantly the earth isn’t alone; it is part of a binary planetary system, orbiting its partner the moon, just as the moon orbits it. So the satellite isn’t just orbiting the earth but also the moon. In addition they all orbit the sun and other planets are pulling on them.

Accurate, synchronized time is also important to GPS calculations. But, that is difficult, because at the speeds a satellite orbits, relativity rears its head. Special and general relativity both play a part in the orbits of a satellite. The clock on the satellite will move slower than one at a tracking station. And as the satellite moves through the orbit it will “emerge” at a slightly different point than where it started its orbit.

So accurately (to the cm or less) figuring out the path a satellite scribes across the sky is difficult in the extreme. What’s amazing is how good they are getting at it. The broadcast ephemerides are much closer to actual positions than they were 10 years ago.

So when do they change the orbit-probably when it shows a real indication of becoming unstable. Not so much just because it moved out of a slot that they want it in. How far is too far out-probably measured in miles and not inches.

But I’m really guessing.

Think Kilometers Yuriy

... tens of meters does not seem like a good guess of mine - since traffic there is not as congested as here, its probably because I am just too fussy (a surveyor) 🙂 .

never studied this question, and your input is very informative, thanks.

FWIW,

Both the IGS and NGA produce a posteriori precise ephemerides. These ephemerides differ from those broadcast by the satellites by being produced from observations of the satellites. In the case of Ultra-Rapid orbits there is extrapolation.

The iGS site for precise orbits is in my previous post. Source for NGA precise ephemerides is http://earth-info.nga.mil/GandG/sathtml/PEexe.html Note that these ephemerides have a five-minute interval. They also differ (if I remember correctly) by explicitly including both positions and velocities. IGS orbits only include positions. Of course velocities are easily computed from positions and times aren't they?

While SV orbits described "merely" using Keplerian elements are satisfactory for planning purposes, the actual broadcast ephemeris contains these elements as well as a number of additional elements. For information about these elements see Guide to GPS Positioning which can be downloaded free from the following site:
http://www.findthatpdf.com/search-81483127-hPDF/download-documents-ln58.pdf.htm BTW, this was the textbook used in my first course on GPS. I especially like its treatment of the broadcast ephemeris elements in the the computation of satellite positions.

For detailed information about the US GPS see its Interface Control Document. The unclassified version is available for free download at: http://www.navcen.uscg.gov/pubs/gps/icd200/icd200cw1234.pdf The ICD includes the statement that during normal operations SVs will have a user range error that is at or below a level required to support a positioning accuracy of 16 meters expressed in spherical error probable.

Another interesting source of information is the description of the operation of the ARC routines in the NGS PAGES software: http://www.ngs.noaa.gov/GRD/GPS/DOC/arc/inp.html

Hope this helps,

DMM

I'm not sure where you got this information but I don't think it is accurate. The SV's definitely do not communicate with each other...

A GPS question - ground Segment

As a side conservation, as I read it, most of the (previous) problem with using GLONASS in RTK solutions was the imprecise location of the sats on the broadcast ephemeris due to lack of ground tracking stations in US hemisphere. I have wondered from time to time why, or if it would be feasible for our GPS system to track other satellite constellations with our ground segment, then add those SVs to the broadcast on the US satellites as informational only. It would effectively double or more the usable on-orbit SVs, and would give the US (and other) users the ability to have more accurate SV locations in different planes. Any SVs with problems could be set to unhealthy and treated the same as an unhealthy US SV. Another country would not be providing the information for the US to upload it would be computed using the same stations and procedures, so there shouldn’t be a security issue. Since they are in the process of upgrading the ground segment it makes sense to me anyway.

A GPS question - ground Segment

Howdy,

FWIW, there is an IGS effort to study how to gather data for all the GNSS. See the Call for Participation at: http://igs.org/mgex/

I recollect hearing plans that the US NGS wanted to establish a network of "fundamental?" sites which would be equipped with equipment that would track at least GPS + GLONASS. I do not know whether this is still a plan, awaits funding or is in progress.

In addition to GPS and GLONASS there are Galileo, Beidou and QZSS.

BTW, the on-line journal Coordinates has a nice discussion of Beidou at: http://mycoordinates.org/beidou/

Cheers,

DMM

> > Yep, the DOD doesn't move the satellite, they guess (really good) where it will be, then they publish the (a posteriori ephemerides) actual positions they were in.
>
> You're probably thinking of the IGS (International GNSS Service) orbits that are computed after the fact.
>
> From the IGS website :
>
> "The International GNSS Service (IGS), formerly the International GPS Service, is a voluntary federation of more than 200 worldwide agencies that pool resources and permanent GPS & GLONASS station data to generate precise GPS & GLONASS products."

You are right it isnt the DOD, I was using the royal "they". As in "they" like making obvious nitpicking comments. Just like in this example sentence the "they" shouldn't be construed to mean the DOD or any other specific person.