Author |
Topic |
|
bo.
Pterodattilo
843 Posts |
Posted - 18/05/2019 : 11:18:12
|
Hi,
Recently I made a flight really close to an airspace border. I found it hard to maintain a proper distance to the border, and in retrospect I realised there was a discrepancy between the infobox value for Horizontal distance and my actual position as plotted on the map.
This was on my Vertica V2 running LK8000 7.1b.
The flight was logged on my android phone using LK8000 running in the background. Replay of this log showed no problems. But replay of the track on the V2 replicated the distance calculation error. The airspace and map files were the same.
During flight, the V2 responded very slowly to manual inputs. For instance, while entering a new QNH value during flight, the numbers appeared a few seconds after clicking on the screen. So I thought perhaps the distance calculation was hampered by limited CPU or memory on the V2. But in SIM mode replay, I could not observe any lagging, but the horizontal distance was still wrong.
Please see the picture in the link below. https://www.dropbox.com/s/ir4ln514usd7n49/2019-05-17%2021.23.47.jpg?dl=0
The map scale indicator can be used for reference.
The airspace to the right (due west) is the one selected in LK8000. Generally the horizontal distance was indicated as around 200 m, even when actually crossing the line.
If needed I can upload the files required if someone will attempt to replicate this problem. Just let me know what you need.
|
Edited by - AlphaLima on 05/06/2019 23:31:54 |
|
brunotl
Pterodattilo
France
1090 Posts |
Posted - 19/05/2019 : 17:33:49
|
igc and airspace files are needed. |
|
|
bo.
Pterodattilo
843 Posts |
|
brunotl
Pterodattilo
France
1090 Posts |
Posted - 20/05/2019 : 10:58:14
|
the problem come from flat map projection used on screen, all platform are concerned...
polygon airspace edge is draw like straight line instead of geodesic line (Arc in flat projection), so if you are in middle of long edge and at hight zoom level, the drawing error can look hudge...
the error you seen also depends to the earth model you used for distance calculation ( always FAI Sphere on WinCE) or WGS84 ( used by default on all other platform )
|
|
|
bo.
Pterodattilo
843 Posts |
Posted - 20/05/2019 : 12:29:28
|
Ok, thanks. Which one is the one to trust, the wgs84 or FAI sphere, and is the infobox or the line on the map more "correct"?
Is this something that can be fixed?
On my phone I have wgs84, and then at least the map and infobox is in agreement. Should I change to wgs84 on the V2 as well? I'm unsure about the implications for other things such as correct turnpoint positions in paragliding comps etc. Any problems there?
Thanks!
|
|
|
brunotl
Pterodattilo
France
1090 Posts |
Posted - 20/05/2019 : 18:36:04
|
hard to answer, my first answer should be none of them... or both..., because it's depends of what is your définition of "correct"
for paragliding comps, you need to choose the same of scoring system used by comps ...
in all case only FAI sphere is available on WinCE platform. |
|
|
bo.
Pterodattilo
843 Posts |
Posted - 20/05/2019 : 20:06:41
|
Hm, ok. Pretty much as expected then...
Do anyone know how the airspace borders are officially defined?
Edit: Answering my own question: ICAO Annex 15 quote: 1.2 Common reference systems for air navigation 1.2.1 Horizontal reference system 1.2.1.1 World Geodetic System — 1984 (WGS-84) shall be used as the horizontal (geodetic) reference system for international air navigation and published aeronautical geographical coordinates (indicating latitude and longitude) shall be expressed in terms of the WGS-84 geodetic reference datum.
And:
European Route Network Improvement Plan PART 1 European Airspace Design Methodology - Guidelines
quote:
The horizontal border of a segregated airspace is specified as a sequence of segments, which can be "straight lines", "along the parallel", "arc of circle", "follow the State boundary". All digital encoding formats (ARINC 4.2.4, AIXM, etc.) are facing common problems with regard to the encoding of this data. The most common ones are mentioned here: #61623; In ARINC 424 straight lines are encoded as “Great Circle” with the Earth being assumed as a sphere. In modern geographical information systems, the most accurate representation of “straight lines” is a geodesic curve on the WGS 84 Ellipsoid. The difference between the two encodings can be significant when calculating the intersection between the intended aircraft trajectory and a segregated area located some 50 NM away, for example.
I would have thought wgs84 was the "original" sphere in earlier versions of LK and XCSoar? The paragliding comps I have attended usually use wgs84 as far as I know. Is it possible to add wgs84 to the PNA version?
Also, it's strange that with the FAI sphere the infobox and the plotted lines in the map are not in agreement. Do they calculate the relative positions differently?
Thanks for looking into this! There is (too much) airspace around locally, so it's pretty essential to be able to trust the information from LK8000. For practical purposes nothing too bad is expected to happen, but then there is the issue of at least trying to stick with the rules, and getting valid flights for the XC league... |
Edited by - bo. on 20/05/2019 20:42:04 |
|
|
brunotl
Pterodattilo
France
1090 Posts |
Posted - 20/05/2019 : 20:58:47
|
WGS84 is used for coordinate (lat,lon) from the beginning, but for calculate distance between 2 coordinate, WGS84 is only available from previous versions.
could be interesting to know how your XC league check the airspace.
the good question could be "what is the correct security margin for avoid to accidentally enter in forbidden airspaces ?" using WGS84 for calculate distance is much expensive compared to using FAI sphere and PNA version is already too slow...
For Comps, you can read this : https://xcmag.com/wp-content/uploads/2017/05/Big-Cylinders-And-Dealing-With-Tolerance-XCMag-July-2017.pdf
|
|
|
bo.
Pterodattilo
843 Posts |
Posted - 20/05/2019 : 21:34:00
|
Ok, thanks. What a mess... But for lower level comps the 0.5 % tolerance does not matter too much I guess. For FAI 1 it will be all wgs 84 ellipsoid from now on.
https://www.fai.org/news/fai-sphere-and-wgs-84-fs2016-r16-and-fs2018-r10-are-released
I still don't understand what numbers to trust on the airspace distances though. Infobox or map, and WGS 84 ellipsoid vs FAI sphere... On Android LK8000 at least I can use the less confusing alternative where the map and infobox infos matches.
The Norwegian XC league has a 50 meter horizontal and vertical (GPS!) tolerance, but I don't know how it is checked. I think there was some software solution to check this. MaxPunkte perhaps? |
Edited by - bo. on 20/05/2019 21:40:42 |
|
|
brunotl
Pterodattilo
France
1090 Posts |
Posted - 20/05/2019 : 21:43:11
|
info box are more precise than map draw for sure ... |
|
|
bo.
Pterodattilo
843 Posts |
Posted - 20/05/2019 : 22:20:13
|
Ok thanks! I will have a look some time then to check if the infobox distances for FAI and WGS are comparable or not. |
|
|
|
Topic |
|