[OSM-talk] Need help with mathematics for runways

[OJW]

On Friday 20 April 2007 19:14, Hakan Tandogan wrote:
> That still leaves me with the problem that I can't convert the X-Plane
> format back to lat/lon of the two end nodes...
>
> Does anyone have a pointer or should I go bury myself in mathematics books?

I've had to do something similar at work (extracting/displaying runway info) 
for another application, so here are a few notes about the experience. They 
may not be relevant if your data source differs from the jeppersen one we 
used (proprietary, don't ask) but I think the fixed-format text files are 
fairly standard.


The runway "position" is a lat/long of the spot where they touch down, where 
the zebra-crossing stripes and runway number are.  There is a "dead 
zone" (unmarked tarmac) behind that line, then there's the runway length in 
front of that line, then an overrun zone at the opposite end of the runway.  
Some data files give you the lengths of those extra zones, which is useful if 
you want to draw the tarmac area as well as just the official runway area.

The width tends to be supplied, and seems reasonably important.  When I did 
it, I created a function MoveTo(LatLong, DistanceMetres, Bearing) that uses 
formulas similar to the ones you found to get from one lat/long to another 
when you tell it the distance and bearing. 
Then:
  RunwayEnd = moveto(Start, length, bearing)
  Corner1 = moveto(Start, width/2, bearing-90)
  Corner2 = moveto(Start, width/2, bearing+90)
  Corner3 = moveto(End, width/2, bearing-90)
  Corner4 = moveto(End, width/2, bearing+90)
  etc.

Bearings tend to be magnetic, but the airport data itself has a magnetic 
variation that you need to add to the runway bearing (presumably so they 
don't need to renumber the runways every time the north magnetic pole moves)

Each physical runway tends to have two entries in the data (e.g. 27R/09L are 
the same bit of tarmac in different directions). This is really useful, 
because if they line-up properly, it tells you that you got the magnetic 
variations correct ;)

There might also be coordinates for a glideslope transmitter (at the side of 
the runway, next to the touchdown point), and for the ILS (at the far end of 
the runway, aligns with the centreline). Similarly, the airport data might 
have coordinates of a control tower, and possibly even some navigation 
beacons.

Regards,

OJW


p.s. just a note: if you can, try to keep as much meta-information as possible 
from the source files (e.g. heights at each end, which airport code each 
runway is associated with, frequencies for each transmitter, that sort of 
thing) because it means that a DAFIF-style file can then be recreated from 
OSM data later. (and as other people have noted, this information could be 
quite a valuable "product" for OSM to offer later)