NMEA Altitude Reading Problems

Narxis · Occasional Visitor Joined: Jun 14, 2004 Posts: 2

Hi all, im writing a GPS application for use on PC / Pocket PC, im reading in NMEA sentences on the serial port from my Fortuna GPSmart BT, all the other figures im reading seem to be pretty accurate except the Altitude, im reading 734m, from a posistion pretty close to see level (OS Maps show between 6-14m above sealevel)

Is there some conversion I need to do to figure out a more realistic Altitude?

Thanx in advance for your help,
Fil

lbendlin · Posted: Mon Jun 14, 2004 6:28 pm Post subject:

Don't even try going there. There is an altitude correction versus the geoid that is transmitted by the sats. Some GPS receivers ignore it, some others factor it in, but the wrong way around, and some others actually get it right. Your program would have to cater for all these eventualities Confused

_________________
Lutz

Report Map Errors here:
TomTom/TeleAtlas NAVTEQ

DavidW · Posted: Tue Jun 15, 2004 12:28 am Post subject:

My understanding was that the satellites don't transmit any corrections - it's up to the GPS receiver to correct altitudes.

GPS altitude is a bit of a nightmare. Not only do you have to deal with ellipsoid to geoid corrections, which some GPSes handle and some don't (and SiRF based receivers seem to do backwards in NMEA mode - that issue is mentioned elsewhere in this Advanced lounge here), but to get from WGS84 altitude to Ordnance Survey altitude you have to carry out a datum transformation.

Add to that that GPS altitude can be rather inaccurate (the horizontal position error is, by design, rather less than the vertical one) and are you sure you want to do this?

David

Narxis · Occasional Visitor Joined: Jun 14, 2004 Posts: 2

Hmm sounds like its not quite as simple as I was hoping it would be Confused

Is the change in Altitude easier to work with? Im thinking that for hikers / riders etc the change in Altitude is likely to be more interesting, perhaps, the user could zero the altitude at there starting point, and see how high they are relative to the start point. Smile

DavidW · Posted: Sun Jun 20, 2004 2:33 pm Post subject:

Especially if you're talking about movement within a ten or twenty mile radius, handling altitude as relative to your start position could work very well. If you find you need absolute altitude handling, you can always come back and rework the code later.

David

UKHABU · Posted: Sun Nov 21, 2004 6:13 pm Post subject:

How products like MemoryMap deal with this is to ignore the height info from the GPSr!
MM has an entirely seperate file for altitude data (after all the actual point you are on the Earth doesn't change). MM then just corellates the latest GPSr live position with the altitude file - the altitude files are *.QED.

My Silva MultiNavigator works differently as the software cross references the GPSr altitude data with an accurate Barometric barometer to come up with a very accurate height - as long as you calibrate it against a known height position eg a Trig point on a mountain top.
_________________
Silva MultiNavigator
GSAK V4.2.2
MemoryMap OS 2004 & V5 Navigator
Navman Pin(Mitac Mio 168)
Navman SmartST Pro V2 & V3
Mitac Mio 558

Skippy · Posted: Sun Nov 21, 2004 11:06 pm Post subject:

The thing to remember is that the altitude problem is with the SIRF based GPS units, other GPS units do actually report the alitude correctly.

You may want to consider an option in your program to ignore the altitude reporting for the broken SIRF GPS units.
_________________
Gone fishing!

DavidW · Posted: Mon Nov 22, 2004 3:50 pm Post subject:

I don't know why I didn't post this earlier. There's some fairly detailed commentary on SiRF based GPS chipsets (mis)handling of altitude in this thread and the thread linked to it in my first post.

David

barryd · Posted: Mon Nov 22, 2004 9:21 pm Post subject:

	PocketGPS
	Web