GPS altitude accuracy???

Go To Last Post
39 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi there,

I would like to know which is the higher altitude accuracy I can get using a gps.

Is there any other way to get altitude?

Thanks.

Michael.

User of:
IAR Embedded Workbench C/C++ Compiler
Altium Designer

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It is my understanding that altitude is generally much lower accuracy than lat and lon. You get something like a few meters x and y but a few 10s of meters in z. I also understand that you can do better on z by taking many readings and average them - like 20 minutes worth. I have heard that this is NOT due to poor receiver operation - just the way whole system works.

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The more expensive Gramins etc. have a presure sensore for doing the altitude because you can get it far more acuratly that way. A friend with a cheaper model logs all his training but when he got 1700 feet during a 2 hour flat water kayak it was clear that gps altitude acuracy was a bit off.

My home made bike computer has a presure sensor for altitude logging. I have quite a bit of software working on the raw numbers which are fairly erratic but I think I end up with good numbers normaly. A presure sensor based method can still be a bit off especialy in some rarer weather situations.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Jim,

Thanks ro the confirmation that was exactly what I knew and you are right the whole systems functionality is responsible for this low accuracy.

ifor,

Could you please tell me a part number of a device like this you are talking about?

Thank you.

Michael.

User of:
IAR Embedded Workbench C/C++ Compiler
Altium Designer

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I've tried GPS altitude both when flying a plane and many times as a passenger on a commercial flight and it is totally pointless (or it was in my Magellan 315A - http://www.avweb.com/news/review... ). If you want an accurate height reading use a calibrated barometer (which is what planes have been doing for decades!)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Look for units that are WAAS enabled. This system improves the vertical accuracy, and was implemented to make GPS usable as an instrument approach method for aircraft.

Tom Pappano
Tulsa, Oklahoma

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Any old 15psi absolute pressure sensor will work... some have amplifiers in them, some dont... but you need at least 12bits to get down to 10ft per bit up high where the pressure is low. 14 or 16 bit external a/d converter wold be better. Seach for 'standard atmosphere' to find an equation for psi to ft

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have heared about SBAS that is usefull for increasing the horizontal position accuracy. Do you know if the WAAS is supported in europe?

Michael

Michael.

User of:
IAR Embedded Workbench C/C++ Compiler
Altium Designer

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
Do you know if the WAAS is supported in europe?

Good question! I don't know the answer to that one.

Tom Pappano
Tulsa, Oklahoma

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have been able to determine reasonably accurate altitudes of fixed locations via GPS by taking altitude readings over several hours/days and averaging them. This obviously doesn't work for moving vehicles.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

EGNOS is live and officially (since early october) working in Europe. It's been in testing-status for a loong while before this. SBAS-using GPS receivers will use this (mine do).

I think your closest monitoring station is in italy, and thus Greece is on the outer edge. Satellite coverage is fine, just the correction data isn't as good as elsewhere.

/Kasper

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

My Magellan is accurate to a few feet of height, going by its readings when I pass altitude signs by the roadside. But it has WAAS.

If I recall correctly, WAAS was introduced to correct the position uncertainty that the military used to impose on civilian users, to make GPS more accurate for aviation purposes. If that's the case I would expect the WAAS-assisted altitude readings to be quite good.

Supposedly they've stopped messing with the position accuracy now. However, when I was traveling recently near China Lake Naval Air Station in California, my GPS informed me I was driving 900 feet below sea level somewhere in the vicinity of Bermuda.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Just to follow up on Jim's answer about poor ALT.

Quote:

I have heard that this is NOT due to poor receiver operation - just the way whole system works.

The reason is simple you need big angles between the satellites to get a good pos. thats fine for LAT and LON but for ALT you would have to use a satellite under you and that one you don't get :)

Jens

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

peret wrote:
My Magellan is accurate to a few feet of height, going by its readings when I pass altitude signs by the roadside. But it has WAAS.

If I recall correctly, WAAS was introduced to correct the position uncertainty that the military used to impose on civilian users, to make GPS more accurate for aviation purposes. If that's the case I would expect the WAAS-assisted altitude readings to be quite good.

Supposedly they've stopped messing with the position accuracy now. However, when I was traveling recently near China Lake Naval Air Station in California, my GPS informed me I was driving 900 feet below sea level somewhere in the vicinity of Bermuda.

The purpose of WAAS (and for that matter EGNOS and MSAS in Japan) is not actually to correct for position, although it does provide some wide-area differential corrections.

It's purpose is similar to what @peret suggested: to allow a civil system to work with an inherently military one.

It's a concept called the integrity of a navigation solution. That is, is the system telling you the truth to within it's accuracy specification.

Integrity isn't accuracy. If a system says I am within 1km of a position and I'm actually 632m away, it's telling me the truth and therefore the solution has integrity, On the other hand, if a position says I am within 10cm (think landing an aircraft!) and I am actually 1m out, then the solution has no integrity, even if it is more accurate that the first solution.

SBAS's primary purpose is to monitor the satellites and the signals that they are broadcasting to determine if the signals are telling the truth. This could be accidental (e.g. a ramp fault), or deliberate (e.g. US in a warzone). Once it determines whether they're good or not, it then broadcasts this information via it's own satellites. And some differential corrections for good measure.

-- Damien

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

There is also another factor affecting height: Geoidal Height vs Ellipsoid Height.

The Geoid is the mean sea level of the earth. This isn't uniform: it depends on the local gravity of the region, which can differ markedly.

The ellipsoid, on the other hand, is a purely theoretical perfect ellipsoid (3-D ellipse) modeled to fit the geoid - GPS uses the so-called WGS84 model. This makes it much easier to give you an approximate height without accounting for the differences in gravity at different locations.

The two are pretty close, but there are significant differences. In Brisbane, it's about a 30-40m difference. Across the world, there's about a +/-100 difference.

The GPGGA string is meant to spit out geoidal height, but some receivers cheat and just spit out ellipsoid height instead. To get geoidal height, you either need a look-up table, or perform a bunch of nasty spherical harmonic calculations...

http://en.wikipedia.org/wiki/Geoid

-- Damien

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

icarus1 wrote:
ifor,

Could you please tell me a part number of a device like this you are talking about?

Thank you.

I am using http://uk.farnell.com/jsp/search... but farnell look to of stoped stocking it. I saw Sparkfun have a breakout board with a SPC1000-D01 http://www.sparkfun.com/commerce... which sounds good although I like the I2C interface on the part I have.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

main error in altitude estimate comes from GDOP errors. To minimize this, ensure the receiver has unobstructed view of the sky. No buildings, trees, terrain.

GDOP errors come from having a partial sky view and poor geometry - to receive from satellites in all sky directions rather than the opposite condition.

Altitude via barometric pressure - of course has to be constantly calibrated using a data source for such as weather varies.

You must tell the receiver which datum you are on. As said above.

WAAS correction data is North America centric. It's sent from geosynchronous (not GPS) satellites such as INMARSAT. As I recall, there is coverage for the Pacific and Atlantic coasts of the US, and most of the land areas. I recall that the EU and Asia have equivalents.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

WAAS/EGNOS/MSAS transmits two types of information interesting to you:
- 'protection information' as already mentioned: Monitoring stations monitor the SVs for health, defined as good range information; Bad range information would be harmful.

- Iono grid. The ionosphere refracts microwaves. The signal path bends some. This changes the path length, which in turn introduces a range error. The bending is different for different frequencies and causes different delays. The military version uses two frequencies, L1 and L2, and computes this error. The new satellites in the constellation, 3 of them I think, also have L2C(ivilian) which will allow not-yet-existing civilian receivers the possibility of computing the error, and thus compensate for it.
Until most of the satellites have this though, it won't be useful.
So, the interim solution is to 'paint' the L1 iono delay using ground stations watching the satellites, and transmit this to the civilian receivers over geostationary satellite.

WAAS/EGNOS/MSAS can not compensate if Selective Availability is turned back on.
DGPS can, but is local in scope.

/Kasper

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The main problem is that 'altitude' is a variable thing. Generally, for aviation issues, you don't give a damn what the absolute height-from-the-centre-of-the-earth is, or even the height above the ground most of the time - you care that your altimeter reads the same as all those other lunatics flying around.

If you're flying at 'flight levels' - it varies by location, but pretty much anywhere above five or ten thousand feet above sea level - then you fly according to a pressure reading calibrated according to a nominal 1013mB and 20C at sea level. The fact that it is almost never that is immaterial; everyone is using the same calibration so everyone is flying at the same level (or offsets between levels). As the temperature changes, or the pressure, you'll move up and down to display the same altitude, but so will everyone else.

When you're approaching an ATC zone, you need to know the height above ground, not the flight level, so they tell you what the reference sea level pressure actually is, calculated from their measured pressure and known altitude, and when you tell your altimeter that pressure reference, you'll get height above sea level; if you just feed in their local pressure reading you'll get height above ground.

As others have mentioned, the gps is already less accurate in the vertical plane than the horizontal, and refers to a notional geoid which only touches the surface of the earth in places - so you need to apply height above ground corrections anyway.

There's a Sema chip which has internal corrections and which is capable of resolving to about 30cm at low altitudes (it's a little worse as you get higher) but it needs a lot of filtering to get a steady reading.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
There's a Sema chip which has internal corrections and which is capable of resolving to about 30cm at low altitudes (it's a little worse as you get higher) but it needs a lot of filtering to get a steady reading.

Neil,

Do you have the part number?

Michael.

User of:
IAR Embedded Workbench C/C++ Compiler
Altium Designer

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

KKP wrote:
. The new satellites in the constellation, 3 of them I think, also have L2C(ivilian) which will allow not-yet-existing civilian receivers the possibility of computing the error, and thus compensate for it.
Until most of the satellites have this though, it won't be useful.

Current high-end receivers ( > $10K, Novatel/Leica/Trimble/Topcon) can also decode the current "military-only" L2 signal using funky cross-correlation methods to solve for the ionospheric errors. Better yet, when receiving real-time corrections from a local base station, they can get to within about 2-5cm horizontally and 5-10cm vertically by resolving the so-called integer ambiguity problem.

There is also a third signal L5 which will allow a bunch of new things to do with GPS.

One batch of satellites L5 have already been launched, but it hasn't exactly been a great success....

http://www.insidegnss.com/node/1478

-- Damien

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Guys I am googling and I can't fint anything related to Sema.

Can anybodey help here???

Michael.

User of:
IAR Embedded Workbench C/C++ Compiler
Altium Designer

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Michael, it's the Intersema MS5534 - http://www.intersema.ch/products/

Make sure you read both the product data sheets and the application notes.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Neil,

Thanks a lot.

Michael.

User of:
IAR Embedded Workbench C/C++ Compiler
Altium Designer

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
I have been able to determine reasonably accurate altitudes of fixed locations via GPS by taking altitude readings over several hours/days and averaging them.

"Good afternoon travelers, this is the pilot. We are getting ready for final approach. We should be on the ground in a day or so."

Regards,
Steve A.

The Board helps those that help themselves.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Koshchi wrote:
"Good afternoon travelers, this is the pilot. We are getting ready for final approach. We should be on the ground in a day or so."

That's why I said fixed locations.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have learned the reasons that make a GPS not beeing accurate, when calculating altitude.

I would like to ask you this: if we have 2 GPS modules from the same manufacturer, working at the same position, then will their altitude (right or wrong - I don't mind) be the same?

Thank you.

Michael.

User of:
IAR Embedded Workbench C/C++ Compiler
Altium Designer

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

When they have chosen the same satellites, assigned them the same weights based on snr/multipath detection magic, and operate with the same almanac(likely) and ephemeris(likely).
And operate at the same incident angle. Don't tilt them too much.

You can't subtract one GPS from another and expect a sensible result. In order to do what you are asking, you need the pseudorange data for all satellites your reference receiver sees, compute the pseurorange errors since you know where you are, and combine that into the navigation solution. This is usually done by feeding the reference receiver's pseudorange error data to the mobile receiver in real time, and let the mobile receiver do the subtraction (this is called DPGS). The alternative (the surveyor's preferred method) is to capture data at both locations, and compute when home in the office.

It's no magic bullet.

/Kasper

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I am asking this because I'd like to find the relative altitude difference between 2 user's, with distance lower than 100m.

So

Quote:
When they have chosen the same satellites, assigned them the same weights based on snr/multipath detection magic, and operate with the same almanac(likely) and ephemeris(likely).
And operate at the same incident angle. Don't tilt them too much.

This must be the key.

Thank you.

Michael.

User of:
IAR Embedded Workbench C/C++ Compiler
Altium Designer

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

I am asking this because I'd like to find the relative altitude difference between 2 user's, with distance lower than 100m.

What was the argument against doing this barometrically - the way almost everyone else measures altitude?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
I am asking this because I'd like to find the relative altitude difference between 2 user's, with distance lower than 100m.

No offence to Icarus, but *this* is the question he should have asked in the first place. No doubt he knows now a lot more about differential GPS than he ever wanted to, but if he'd started with this question, we'd have said...

Go forth and buy an altimeter. Actually, buy two. A standard para/hang-glider vario runs from around a hundred pounds to as much as you want to spend.

At a hundred meter's horizontal separation, you're in the same base pressure system. The vertical displacement on a digital unit will be within a meter or so.

Or even, at that distance, if you can see each other, borrow a surveyor. Two points a known distance and height apart, two readings of angle to the third point, a little trigonometry, and Robert is your aunt's husband.

It all depends on how much accuracy you require, how fast you need the answer, how fast things are moving. But the most hi-tech answer is often not the best.

If you choose to build a vario, I have some working code for the Intersema chip I referred you to.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I've been using GPS to measure three points on an object and can do it to centimeter resolution. The GPS's in question are $15000 each though with the antennas another $1000 a pop. I get a X,Y,Z vector between each receiver and it's master. As KKP mentions, using cheapo receivers you could get the raw observations then post-process on your PC as it requires an amount of computation. I doubt this is what you want to do. Seems baro is the simplest solution.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I got you guys. Maybe I might post that I needed the relative altitude and not the actual, but it was an erroneous omission.

Thank you all for your time.

Michael.

User of:
IAR Embedded Workbench C/C++ Compiler
Altium Designer

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
As KKP mentions, using cheapo receivers you could get the raw observations then post-process on your PC as it requires an amount of computation. I doubt this is what you want to do. Seems baro is the simplest solution.

The phase noise on survey grade receivers (your $15K receivers) are an order of magnitude less than (for example) a $100 U-Blox, not to mention that you have dual frequency measurements. While possible to process the lower-grade receivers, the high quality receivers make this job much sooooo much easier. We have problems resolving integers reliably using a U-Blox but can fairly easily do it with a good receiver.

-- Damien

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Novatel have this software:

http://www.novatel.com/products/...

At the price of this software , one would think you wouldn't waste your time with el -cheapo receivers, nevertheless, it can make L1 only receivers do some nifty 'tricks'. Unfortunately it's acquisition time is a little slow for the app so we went for the full-ticket items. Nevertheless, it can do some of the stuff you get with L1/2 and can sometimes work out more economical than going to a L1/2 solution.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
Novatel have this software:

http://www.novatel.com/products/...

NovAtel stuff is truly drool-worthy, and the post-processing allows for a forwards-backwards type solution, but I'd hate to think what you paid for the licence! We're getting a little off-topic...

-- Damien

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Actually what I meant to post was:

http://www.novatel.com/products/...

It was less than a L1/L2 receiver.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

OT: I can see where altitude is important to a user named "Icarus". We wouldn't want him to get too close to the sun, now, would we?
http://en.wikipedia.org/wiki/Icarus

Lee

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

That's right.

Michael.

User of:
IAR Embedded Workbench C/C++ Compiler
Altium Designer