GPS & geosynch'd satellites - do they "wander"

Factual Questions
1

I think I understand triangulation. I think I understand how the GPS system works. I may be wrong on both counts.

I thought that to triangulate you needed at least three known, fixed locations. Barring large-scale geological change, a radio transmitter on our planet’s surface remains fixed.

I know GPS satellites are in geosynchronous orbit. I know what “geosynchronous” means – it means fixed over a specific point on the planet’s surface.

HOWEVER!!! It is my understanding that GPS is very accurate – to within a meter (or couple of meters). To achieve that degree of accuracy, repeatedly and over time, wouldn’t the satellite need to remain fixed over the exact same spot, repeatedly and over time? If it “wandered”, wouldn’t the readings on the ground be thrown off?

For example, I thought of a weather satellite hanging geosynchronously over the Atlantic ocean. If it “wandered” five meters a year, in 10 years it would be 50 meters out of place. Well… is that a big deal? The Atlantic is huge, space is huge, and 50 meters one way or the other probably won’t make any difference when you’re snapping pics of a hurricane hundreds of km across.

But… if a GPS satellite “wandered” 5 meters a year, in 10 years the readings on the ground would be significantly off, more than the couple of meters of accuracy it currently achieves.

So… (1) is my understanding, logic, or thought process totally off base? (2) how fixed are GPS satellites? (3) do geosynchronous satellites “wander”? (4) if number (3) is true, how big a problem is that, and how is it addressed?

Thanks.

2

Even geosynchronous satellites can wander. But they don’t keep miving in the same direction over time. For reasons that would take a good semester of orbital mechanics to understand, any wandering they do would take place in a roughly “figure-8” shape. This ground track may even precess a little, kind of rotating about its own center over time. But its average position is always over the same spot.

Furthermore, this wandering can be easily predicted and compensated for. Orbital mechanics is nothing if not predictable.

3

Actually, GPS is much more accurate than that. The hand-held instruments sold at sporting shops give instantaneous readings that are only accurate to a few meters. But other instruments are accurate to one millimeter. Land surveyors and aerial mapmakers use these (expensive!) instruments.Mounted on a stationary tripod, it takes about a half an hour to average measurements from many satellites .

I dont know any technical details, but somehow the satellites “know” if they wander, and the GPS instruments take this into account when calculating their position.

4

GPS satellites are not in geosynchronous orbits - they’re much lower, and zip overhead. It’s a complex system, but you need four satellites to tell you your position. If you knew the distance to each satellite, it would take only three, but a GPS receiver doesn’t know that. It only knows that at time T it received a satellites signal that should have been transmitted at T-t. But since the receiver doesn’t know the exact time it is now, it takes another data point, a signal from another satellite, to have enough information to figure it all out.

5

GPS satellites are NOT geosynchronous. They move in relation to the earth’s surface. If you go to this site: http://science.nasa.gov/RealTime/JTrack/3d/JTrack3D.html You can see a representation of all the satellites. You can click on one and it’s designation and orbit will be displayed. You will see that the GPS constelation is well inside the orbits of the geosynchronous satellites (Clarke Belt).

The GPS system works because the satellites are in very stable orbets and the ground units are programmed with the information predicting where each one will be far into the future. Based on the predicted position and reception of their transmitted signals, the receivers display your position

6

Wrong on every count, my friend.

GPS satellites are not geoscynchronous, and you do not need fixed positions for triangulation.

The satellites have their predicted orbits, and your receiver uses the positions calculated from the day and time and the signal delay of the time stamp in the data packet sent from the satellites to figure out where you are.

To figure out where you are, your receiver first assumes that it knows the correct date and time and that it has a half assed guess of your position from the last time it was used. It uses this data to figure out which satellites should be overhead - and then proceeds to attempt to decode data packets from those satellites. If that fails - either because the date and time or wrong or the old position is completely wrong, then the receiver does a scan of all the satellites to try to find one that it can get the correct date and time for. Once it has that, then it at least knows the hemisphere it is in and also has a proper time and date so that it can go back to looking for only the ones that should be overhead.

Once the receiver has enough data, it figures the exact position of all of the satellites it is using and then figures your position from there. The normal data packets are a time stamp that begins at precisely that instant in time for which the stamp is valid. If the time stamp says “1200 greenwich,” then the first data pulse began at precisely that time.

The clocks on all of the satellites are synchronized to some utterly fantastic value - nanosecond range, if I remember correctly.

The clock in the reciever sychronizes itself with that of the satellites. To tell how far from a satellite it is, it measures how late the time stamp is when it arrives. It started at precisely 1200 (in our example) and gets to the receiver at 1200 plus some odd number of microseconds. Since the speed of light is known, you can calculate distance easily, and use this for your triangulation.

The clocks wander a bit, and the satellites don’t precisely follow their predicted paths. for these reasons, the US goverment runs several ground stations that monitor the satellites. Variations in orbit and time get noted and passed up to all the satellites periodically. The satellites pass this data on to the receivers. It takes about twenty minutes for a complete “ephemeris” update to come down. For this reason, a GPS receiver that has been out of the loop for a while (shut off and not receiving for a few days) is a bit inaccurate until it has managed to get a complete ephemeris. This data tells it how to correct the orbital data and the time stamps for all of the satellites.

The satellites can also be individually shut down. This happens on those occasions when a satellite starts giving out really bad data. Sometimes this can be fixed from the ground and the satellite gets reactivated. It has also happened that satellites have been permanently taken out of service for some failure that can’t be corrected via remote control.

Awful complicated, ain’t it? Sort of makes you appreciate what goes on in that little receiver box a little more, don’t it?

7

That’s the word I was trying to remember!

Excellant explaination BTW.

8

And those units are not strictly straight GPS. They use a lot of tricks on the receive side and some pretty hefty postprocessing back in the office to get that accuracy.

One of the things that they do in post processing it to have a receiver set up in a known position relatively close to the survey area. The position of this antenna is surveyed most carefully by standard surveying methods. They then use the diffence between the know position and the calculated postion to give more detail on the ephemeris data. When you get into accuracy levels like this, even the atmospheric conditions play a role so that you need correction data from the area you are in - not just from where ever the ground stations are.

Another thing they do (in the receiver) is to synchronize the clock in the receiver to the higher speed signal that the military uses. The civilian receivers you can buy use a 1MHz signal with an encoded data stream - the government lets anyone know how to decode this signal. The military signal runs at 10MHz, and is mroe precise by that factor of ten - but only the military can have units that can decode that signal.

The civilian surveying units can’t decode that military signal, but they can use it as a timebase for their own clocks. Since that 10MHz signal is synched to the clocks on the satellite, it is much more accurate than any oscillator that a civilian would want to pay for.

Normal civilian GPS receivers sync their own clocks to the 1MHz signal that they are decoding. The surveyor’s equipment has an set of extra receive circuits that pick up the pulses in the 10MHz signal. This makes them more complex and hence more expensive.

The other thing they do is collect boat loads of information. The time stamps come out once per second. Collecting data for 30 minutes gets you 1800 data points that can be post processed with the base station data get very good results. Since the field unit and the base unit log their measurements using the timestamps from the GPS satellites, it is no trick at all to post process any given data point from the field with the proper correction data from the base unit.

9

Thankee, thankee!

The company I used to work for did some work on GPS tracking equipment to be attached to ordinary portable two-way radios.

The project sort of died a long protracted death, but that’s where my experience comes from - trying to shoe horn one into a little tiny box and make it run on the tiny bit of current most radios deliver to accessories. We finally got it to work, but it turns out nobody really wants to have it. They all say “great” when you tell them you can use it to find their injured fireman whose out fighting forest fires - but then the unions get wind of it and decide that it could also be used to see who is spending too much time goofing off in the donut shop, and that kills it.

10

How do you like the Garmin Rhino?

I’m a Ham and we use our GPS receivers to feed our APRS (automatic position reporting system) equipment.

11

nitpick: The ephemeris that each satellite sends about itself only takes 30 seconds. The almanac that gives the ephemeris for all the satellites (to aid in searching) takes the 12.5 minutes. But a complete sky search without the almanac (“cold start”) shouldn’t take more than a few minutes, at which time the satellite positions you have are at full accuracy.

12

GaryM: Interesting Java! I’ve been clicking at random, and I’ve found some interesting birds. Chandra, for one, and Vanguard, not to mention the space station. I’ve noticed that GPS birds tend to be off the ecliptic (for obvious reasons), and that the Earth has developed quite the little ring system. :slight_smile:

Also, I found (and promptly lost) a SKYNET bird. Interestingly, SKYNET is a real military project: “United Kingdom Ministry of Defence Geosynchronous Military Communications Satellite Programme”. One can only wonder… :wink:

13

It may be a nitpick, but GPS receivers don’t work by triangualtion. I came up with the following definition for this word:

Triangulation: Finding a position by means of bearings from two fixed points a known distance apart

GPS receivers don’t calculate bearings. As Mort F. noted, their calculations are based on the time a received signal took to travel from the originating satellite, and information about the exact paths of the satellites that effectively allows calculation of the position the satellite was in when the signal was sent.

As has been noted, the satellites are not geosynchronous. Being lower, they orbit in something like 12 hours, which makes it possible to control them from a single ground station - signals are sent to them as they pass by. It also means they need less power to deliver sufficiently strong signals to the earth’s surface.

The fact that the satellites move makes the computations more complicated. But as long as the movement is known and predictable, accuracy need not suffer.

14

As has been noted, GPS satellites are not geo synchronous and normal hand held units are prone to errors due to various foctors. If your GPS supports WAAS, you both lock into a geosychronous satellite, and get a batter result (<3 metres). Only workin the USA though.

15

Following what Xema said:
The word for the day is trilateration: position determination using the distance to known objects.

Following what Tapioca Dextrin said:
In Europe, you can use EGNOS.

16

Nifty, but not what we were doing.

We were working on a speaker mic with a GPS receiver that you connect to any portable radio.

We were tying into an existing system for fleet management that handles up to 65000 units, and plots positions for them on screen using mapping software in the communications center.

The cool thing about our system was that it was extremely low power - it only drew about 80mA, and that only for a few seconds every minute or so.

The othe really cool thing was the communications protocol that was used over the air. At 12dB SINAD, the messages were guaranteed to get over 90% of the time - and 50% at 0dB SINAD.

A truly ass-kicking system, and none our customers wanted to have it - even though we built it based on suggestions from them.