Closed joe-saronic closed 4 months ago
@phayes Could you have a look?
Hi @joe-saronic, thanks for the great bug report!
@urschrei, can you assign this ticket to me please? I'm not sure that I'll be able to get too deep into it today, but I'll have time next week.
Thank you! Done.
On a more careful reading of the documentation, I am beginning to think that what I was actually looking for is the Haversine destination. It looks like geodesic destination actually respects the non-spherical model, which correctly accounts for the rotation of the geodesic as it goes around the Earth.
For anyone else following up on this ticket, travelling around a geodesic at 45 degrees should look something like this (note that this is exaggerated in comparison to earth, earth has f = 1/298.25) :
Note that the only closed geodesic routes are where we are travelling at either 90 degrees (north-south), or 0 degrees (east-west) around the ellipsoid. Travelling these directions should result in a closed cyclic route.
My expectation is that repeatedly traversing a great circle with
geodesic_destination
will yield identical lat/lon/bearings up to floating point roundoff or so. I have made a small script to show some inconsistencies:Here is a python script to plot the results against each other:
The bearing may be a bit of a red herring in this case. However, the lat-lon relationship should in theory produce identical results with multiple passes, yet it does not appear to:
Zooming shows that the circle is rotating approximately one degree in longitude with every full rotation. This does not seem like something that is ascribable simply to round-off error.