Re: [Paparazzi-devel] Re: Paparazzi-devel Digest, Vol 61, Issue 16

[Pascal Brisset]

Christophe De Wagter wrote:
> I believe that within a UTM zone the NED coordinates and UTM 
> coordinates are only translated (infinite UTM zone).

Unfortunately it is not the case. If you set a reference close to 
Toulouse and compare NED (or LTP for Local Tangent Plan) to UTM 
coordinates for a point 1000m away, there is a difference of 20m. 
Moreover, while the LTP is plan (!), the UTM projection is done on a 
cylinder wrapped around the earth, so the altitudes do not coincide in 
the two systems: 10km from the reference point, there is already a 
difference of about 10m.
So, if the NED (LTP) coordinates are perfect for a "local" flight, they 
cannot replace UTM, at least in a simple way, for flights some 
kilometers around the reference point.

Altitude is really a problem when you want to use ECEF since there is 
no simple transformation from ECEF to LLH with the usual "H". While 
there exists an analytic formula to compute a height above the ellipsoid 
(the WGS84 model which tries to fit the earth), the computation of the 
height above the geoid (a model of a medium sea level) requires a 
database. I'm curious to know how the GPS receivers compute this height:
 http://en.wikipedia.org/wiki/Geodetic_system
 http://en.wikipedia.org/wiki/Geoid

--Pascal

PS: Some test code is available in sw/lib/ocaml/test/test_latlong.ml to 
be compiled with
    ocamlc -I .. unix.cma str.cma latlong.cmo test_latlong.ml