GPS accuracy

[Adi007]

I've been asked a few questions on the accuracy of GPS. Well, I believe that a picture speaks a thousand words so I hooked my Garmin Vista to my iPAQ 3950 Pocket PC and used Memory Map plot my travels on a 1:50,000 OS map. Plots are taken every 15 meters or 10 seconds automatically, whichever comes first.

Here are the results:
----------
First, accuracy on major roads. The accuracy is good enough to show which side of the duel carriageway I'm on!

----------
Here is another ... remember, this is plotted data and there is no "lock to roads" or anything enabled (there isn't such a feature on Memory Map:

----------
Closeup of the accuracy ... this is overtaking a vehicle!

----------
Performance in also exceptional on small roads:

----------
This shot shows a "vector arrow" that shows the distance that will be traveled in 10 mins (customizable). This is really handy on motorways!

----------

 

[ESpy]

I've seen the Garmin EMap get decidedly confused when roads fork and there is tree cover (with snap to roads set) - it'll merrily go down the road I haven't taken, which can upset a navigator quite a lot. It loses lock and tries to plot a track derived from original heading & speed coupled with the snap on.

Apart from that, the clock jitter claims accuracies in the order of 20ft, which is an improvement on the old SA being switched on. We once spent about 2 hours doing a search for a wreck that the GPS said we were right over.

 

[Leon]

Ive used a Garmin Etrex out and about for a while and found it pretty darn good down to 10 metres. When I took it on holiday to Boston to use in the city however, ...useless. Couldn't get a position due to all the skyscrapers :-? (I allowed for the fact that we'd travelled more than 500miles from the area of normal use, but given all the time in the world, it still couldn't [or wouldn't :evil:] get a fix).
How does GPS in cars etc overcome this problem; more power??[/i]

 

[Adi007]

Ive used a Garmin Etrex out and about for a while and found it pretty darn good down to 10 metres. When I took it on holiday to Boston to use in the city however, ...useless. Couldn't get a position due to all the skyscrapers :-? (I allowed for the fact that we'd travelled more than 500miles from the area of normal use, but given all the time in the world, it still couldn't [or wouldn't :evil:] get a fix).
How does GPS in cars etc overcome this problem; more power??[/i]

The problem with GPS is areas where there are tall buildings is an error called multipath error that results from the GPS signals are reflected off the building walls. If your GPS was off and you moved a significant distance )say over 100 - 250 miles), you need to allow the GPS to re-download the almanac and reposition itself.

the other problem is that if you are in a car with metalized windscreens you'd need an external antenna because the signal can't get through them.

 

[Leon]

Thanks Adi,
I guess it must have been tne multipath problem. I was aware of the problem of travelling over a significant distance with the unit switched off and after allowing extra time it did give me a position out in the open (and Quincy Market, oddly)...just in amongst the buildings was a problem.
I didn't try it in a car, just got me wondering how cars even with appropriately fitted systems deal with the tall bildings in cities like Boston, London?
More pertinent to my little Etrex though, I wondered how other similar handheld units cope, given that some of them come with city street maps programmed in, yet don't appear to have any better system fore actually receiving the signal :-? I refer to walking with them, rather than sitting in a car.
Thanks again,

 

[Adi007]

If I'm somewhere where I get a poor signal with my GPS (trees or buildings) I use something called a re-radiating antenna which is a more powerful 12 V powered antenna to pick up the signals from the satellites and reradiate it for the GPS antenna - i find that woks well.
http://www.freshtracksmaps.com/antenna-vortech.htm
Outdoors however, pretty much all the modern hand held devices use antennas that are of similar performance.

 

[Leon]

Thanks Adi :notworthy

 

[leon-1]

Adi, from what I can recall nearly all GPS work from the satellite network that is effectively controlled by the military. Didn't the military at one stage use a specific codec so that the GPS was not quite as accurate for the civilian market?

I know that this would hinder people in boats and shipping lanes, however the accuracy for these units was increased by using a beacon receiver which was used to cross correllate your location.

I seem to remember the people at Garmin saying that they would not gaurentee the accuracy of a handheld unit to be less than 10 metres unless it was attached to a beacon receiver, in which case accuracy would be gaurenteed down to 5 metres (on this they also stated that normally the unit was accurate to about 1.5metres).

Do you know if this has changed?

 

[Adi007]

GPS satellites are indeed military satellites, controlled by the US Space Command. So, if they wanted to they could turn the signal off at any point ... however, given the widespread usage of GPS nowadays this seems unlikely. Add to this that Europe is planning on launching GPS satellites of it's own.

GPS satellites (called SV - short for Space Vehicles) emit two signals at slightly different frequency (1575.42 MHz and 1227.60 MHz).

• P Code - Precision code. The signal that the military uses. If you think of the signal as a ruler extending from the SV, the "length" of the ruler is approximately one week multiplied by the speed of light, and each of the "graduation" on the ruler is 30m apart.
  CA Code - Coarse acquisition. Using the ruler analogy, this can be thought of as a number of physical rulers extending from the satellite to the receiver. The length of each of the ruler is approximately 300km, and each of the marks on the ruler is 300m apart.

The difference between the two is subtle but needless to say that the P code is slightly more accurate that CA code and P code can also be encrypted. Both are still in use (the P code is used in guided munitions and ships and aircraft while most "on the ground" units now use CA because the receivers are cheaper.

Prior to 2000, the CA signal had an introduced error called "Selective Availability" that deliberately reduced the accuracy of the CA signal to something like 75 - 100m. To try to get around this many units had a "position averaging" feature that would give you an accuracy close to 25m. Not bad but still not excellent!

This error has now been removed and you can expect accuracy from a civilian unit of about 10m in good conditions (away from trees and buildings - the GPS signal is absorbed by the chlorophyll in leaves). To improve accuracy a system called DGPS (differential GPS) was introduced and this used two antennas a known fixed distance apart to collect the GPS signals from the SVs and get even better position accuracy. This was expensive.

Enter a system known as WAAS - Wide Area Augmentation System. This is the beacon system you mentioned. There can be ground based beacons or satellite. Europe is now served by a WAAS signal from a satellite (EGNOS) and if you buy a GPS receiver that is WAAS compliant you can get accuracies of between 5 and 1.5m.

One thing to remeber with GPS is that nothing is guaranteed - batteries can die, signals can be lost, the terrain can hinder the signals ... so always have a backup!

 

[stevec]

i seem to reacll that it was bill clinton that repealed the selective availibility act thing, good old bill!!!!! :shock:

 

[Adi007]

i seem to reacll that it was bill clinton that repealed the selective availibility act thing, good old bill!!!!! :shock:

It was indeed!

 

[leon-1]

Cheers for clearing that one up guys, I hadn't kept up with it quite as much as I would have liked (despite the fact that I work in the telecommunications industry now).

 

[Mikey P]

Pedant's Corner:

P code is always encrypted and is known as P(Y) code. If you have done any reading about GPS or work in the industry (Adi007 - you certainly know your onions! Do you work for Garmin/Magellan?), you'll see that GPS is an incredible system and we found it absolutely invaluable in the Gulf last year.

The military are supposed to use P(Y) code receivers, the handheld versions are often known as PLGRs ('pluggers'), but the PLGRs generally take longer to acquire, there aren't that many around, and they lack most of the user-friendly features found in the commercial models (like the excellent Garmin Vista). Even worse, you need the crypto fill to decode the P(Y) code signal which means that, if you lose it, you're right in the clag!

In my experience, just about everyone I orked with bought their own CA-code civvy GPS and the first thing you did when meeting another unit in was to swap position details on hazards, mines, ambush areas, UXO, etc. I would honestly say that they have been life savers.

There are problems with everyone running about with civilian GPS receivers (which I won't go into here) but, as long as they are used as an aid to map and compass work, I think they are excellent.

My only comment on using GPS in cars is that they generally have a better externally-mounted antenna and the elevation angle to the satellites remains constant - if you have a hand-held system, you may hold the antenna at a non-optimum angle or even shield it from the SV constellation with your body.

And, yeah, well done Bill Clinton!

Maybe Adi007 should do a tutorial on GPS usage?

 

[Adi007]

There are two other advantages to using P(Y) encrypted code ...

1 - Because it's encrypted you can't have someone sending out spoofing code to put you off course.

2 - The code can be used to determine distortion because of the atmosphere (in particular, the ionosphere), therefore making the receiver more accurate.

Apart from the fact that P-code is inherently more accurate than C/A-code, civilian units are far better than most military units in terms of speed or acquisition of signal and features. Signal augmentation (such as WAAS), have dramatically improved the accuracy of GPS navigation.

No Mickey P, I don't work for either Garmin or Magellan. :-D

 

[AntarcticFire]

I've been wondering what EGNOS is for quite a while... we did development
for it at work and a load of the equipment in the comms room
is labelled up for EGNOS... I just never knew what is was

 

[Mikey P]

I seem to remember that the US was looking at additional C/A channels in the next iteration of GPS - this would allow you to correct for atmospheric distortion as in P(Y) code. Is this the case or did I dream it?