Author Topic: Pushback at Axonos YPAD stand 20L completely wrong  (Read 2134 times)

Fiorentoni

  • Full Member
  • ***
  • Posts: 129
Pushback at Axonos YPAD stand 20L completely wrong
« on: August 19, 2022, 08:38:27 pm »
As in the title: I chose "right" and what GSX did is push me 360 degrees around so I ended up with my tail in the terminal and the nose versus the taxiway. That's where pushback ended.

virtuali

  • Administrator
  • Hero Member
  • *****
  • Posts: 51439
    • VIRTUALI Sagl
Re: Pushback at Axonos YPAD stand 20L completely wrong
« Reply #1 on: August 19, 2022, 08:59:33 pm »
Which is why there's a custom pushback feature, to sort out special cases, issues with the scenery itself, etc.

Fiorentoni

  • Full Member
  • ***
  • Posts: 129
Re: Pushback at Axonos YPAD stand 20L completely wrong
« Reply #2 on: August 19, 2022, 09:04:00 pm »
But how can I know that before starting the pushback? Can't do that stuff on Vatsim.
Also why does this happen? I thought GSX reads the taxilines from the scenery?
Interestingly when I used the quick edit mode the suggested pushback route was *correct* (not 360 degrees)....

virtuali

  • Administrator
  • Hero Member
  • *****
  • Posts: 51439
    • VIRTUALI Sagl
Re: Pushback at Axonos YPAD stand 20L completely wrong
« Reply #3 on: August 19, 2022, 09:11:27 pm »
But how can I know that before starting the pushback? Can't do that stuff on Vatsim.

Sure you can, just always use QuickEdit, which you only need to do it once, because after that is saved for that gate. The Quick edit path preview it's a simulation of the actual IK system of the truck+airplane system, it is doing the same calculations as if it was actually pushing, so you can be sure the path you see there ill be the one what will be performed, with that truck and that airplane.

Quote
Also why does this happen? I thought GSX reads the taxilines from the scenery?

Of course it does, which works in most of the cases, but that doesn't mean it's always possible to find the "best" position, from a real life point of view. The mathematically solvable, might not be the best looking.

Quote
Interestingly when I used the quick edit mode the suggested pushback route was *correct* (not 360 degrees)....

The QuickEdit, initially, doesn't suggest anytying ( if you never used it on that gate before ), it just start some distance and some angle away from the airplane. so, the first thing you should to, is to place the airplane final point. Then, usually, press NumPad5 to place two waypoints at the start/end of the path, and their main function is to be sure the end path (particularly) ends up with the tow truck aligned on the taxi. Then, if needed, you might want to tweak the waypoints a but, but it's hardly necessary.

Fiorentoni

  • Full Member
  • ***
  • Posts: 129
Re: Pushback at Axonos YPAD stand 20L completely wrong
« Reply #4 on: August 19, 2022, 09:19:14 pm »
Okay thanks, but this might be of interest to you, that the preview with quick edit showed a correct pushback while the actual "to the right" pushback did spin me around 180 degrees (not 360, sorry). That's surely not mathematically correct to quickly throw my aircraft around and say "pushback complete". I was still next to the jetway at the end of that "pushback", just the other way around:

          Tug
           
           ^
           *
      *******      JET
           *            WAY
------TERMINALBUILDING--------------------------

So it did either two different calculations (one of them wrongly) when I clicked quick edit and after that "to the right" OR the actual execution was not done via the calculated route for some reason.

virtuali

  • Administrator
  • Hero Member
  • *****
  • Posts: 51439
    • VIRTUALI Sagl
Re: Pushback at Axonos YPAD stand 20L completely wrong
« Reply #5 on: August 19, 2022, 10:26:52 pm »
the preview with quick edit showed a correct pushback

It was just for pure chance. In fact, the very fact the generic QuickEdit start position happened to be in a better place, indicates that spot is not very common so, what is usually "wrong" ( the default QuickEdit position is not very useful in normal spots, as it is without any editing ), because "right", when used in a different context.

Quote
while the actual "to the right" pushback did spin me around 180 degrees (not 360, sorry). That's surely not mathematically correct to quickly throw my aircraft around and say "pushback complete".

That's exactly the opposite: because it MUST be mathematically correct, it would result in a bad *physical* representation, because when you add realistic mechanical constraints to something that would work from a purely mathematical point, in that situation it would result in a locked truck, because the IK couldn't be solve, similar to when jetways don't move if they cannot solve for the door.

As I've said, most of the times the automatic pushback can figure out a good solution, but this is not valid everywhere, and it's a combination of so many factors all added together: the length of the tow wheels axles, the towbar length if there is one, the distance between the front and rear gears of the airplane, the existing nodes in the airport and how they are placed, if there are nodes closer than the distance of the whole truck+airplane system, etc, etc. ALL of this is taken into consideration, but without mechanical constrains so, you might see 180s, but no lockups.

And again, for THOSE situation, it's where you need QuickEdit.

Fiorentoni

  • Full Member
  • ***
  • Posts: 129
Re: Pushback at Axonos YPAD stand 20L completely wrong
« Reply #6 on: August 19, 2022, 10:39:24 pm »
the preview with quick edit showed a correct pushback

It was just for pure chance. In fact, the very fact the generic QuickEdit start position happened to be in a better place, indicates that spot is not very common so, what is usually "wrong" ( the default QuickEdit position is not very useful in normal spots, as it is without any editing ), because "right", when used in a different context.

Quote
while the actual "to the right" pushback did spin me around 180 degrees (not 360, sorry). That's surely not mathematically correct to quickly throw my aircraft around and say "pushback complete".

That's exactly the opposite: because it MUST be mathematically correct, it would result in a bad *physical* representation, because when you add realistic mechanical constraints to something that would work from a purely mathematical point, in that situation it would result in a locked truck, because the IK couldn't be solve, similar to when jetways don't move if they cannot solve for the door.

As I've said, most of the times the automatic pushback can figure out a good solution, but this is not valid everywhere, and it's a combination of so many factors all added together: the length of the tow wheels axles, the towbar length if there is one, the distance between the front and rear gears of the airplane, the existing nodes in the airport and how they are placed, if there are nodes closer than the distance of the whole truck+airplane system, etc, etc. ALL of this is taken into consideration, but without mechanical constrains so, you might see 180s, but no lockups.

And again, for THOSE situation, it's where you need QuickEdit.

Alright thank you, that explains quite well what occurred and why. Keep up the good work.