Project JBT B1200 Aircraft Tug

[tl8]

As it seems there is some interest in this project I will outline my ideas and problems in this thread. I will add more later (Along with images).

Attached is a zip file which contains a (probably) recent version of the add-on so far. This isn't even in Alpha stage so I haven't put things in the right folder yet.

The real thing:
http://www.jbtaerotech.com/en/solutions/equipment/aircraft-tow-tractor-b1200.aspx

The mesh used is from the Shuttle Tug add-on and is only used to get the code working.

Edit: I have now uploaded some images. The most interesting one would be the first one which shows the internal view of the tug.

Also there are a few external shots showing the wheel animation and the XR5 attached to the tug (not moving)

I have also updated the zip file to the latest dll (2010P1).

For those inclined I am happy to share the code.

Edit2: Uploaded again with with file tree also includes a more complete manual. Note C++ 2010 EE is probably required.

 

[Scruce]

Good idea, does it work with an attachment or something?

 

[DaveS]

Just tested it here and I get a CTD upon launching the included test scenario. That is after I have deleted the XR-5 entries and installed the msvcr100d.dll file.

 

[C3PO]

CTD here too. Did you use subdirectories for mesh or textures??

 

[tl8]

Ok, I might recompile as release and put it in a folder tree.

 

[DaveS]

Just tried the updated version, still CTDs. I'm going to install VC++2010 Express and see if that changes anything.

---------- Post added at 08:02 PM ---------- Previous post was at 02:18 PM ----------

Installed VC++2010 Express Edition, still CTDs.

 

[orb]

Just tested it - not crashing for me, though when I attach to it the XR5 from the included scenario and then accelerate, the tug starts spinning, and either the XR5 disappears, or there are added some plasma effects, and there is displayed this debug string (first there were some numbers, which were growing):

theta1 1.#QNAN0 1.#QNAN0 1.#QNAN0 1.#QNAN0 1.#QNAN0

I don't know what it is, but I may check it later by looking at the included source, and running it in debugger.

 

[DaveS]

Just tested it - not crashing for me.

Maybe it is because I have removed the XR5 from the scenario and have nothing but the tug in it? Is the XR5 a requirement?

 

[orb]

Maybe it is because I have removed the XR5 from the scenario and have nothing but the tug in it? Is the XR5 a requirement?

No it isn't. I first tested it without the XR5, because I hadn't got it in my testing installation, but later I copied the tug to an installation having XR5, and tested it against it. No CTD in either of them.

 

[Spacethingy]

Nice one tl8. :thumbup:

This'll go well with the new WIN that's in the pipeline!

 

[tl8]

Just tested it - not crashing for me, though when I attach to it the XR5 from the included scenario and then accelerate, the tug starts spinning, and either the XR5 disappears, or there are added some plasma effects, and there is displayed this debug string (first there were some numbers, which were growing):

theta1 1.#QNAN0 1.#QNAN0 1.#QNAN0 1.#QNAN0 1.#QNAN0

I don't know what it is, but I may check it later by looking at the included source, and running it in debugger.

Yes, that is why the project is stalled.

Unless you understand Kinematics and Dynamics, you will have trouble fixing it.

The debug string shows the values of some variables used in the rotation of the vessel (for debugging), one of the problems could be the mass of the XR5 but from memory I have tested that and it doesn't help.

Edit: Sorry DaveS other than that I have no idea why it doesn't work. Are you using 2010P1?

 

[orb]

The included source is probably from some earlier stage of development, as after attaching a vessel to the tug, this bug doesn't exist, but the connection between tug and a vessel is rigid. There is only the same bug with wheels being animated off-axis when I turn and move.

 

[tl8]

Try this then

http://dl.dropbox.com/u/3550991/JBT_B1200.cpp

 

[tl8]

A little progress.

After a bit more work, I have found that the system I am using is not right.
I have now found a possible solution which is to use a rigid body physics engine.

I am deciding between newton or bullet but bullet seems to be easier to understand. Integration shouldn't be too hard as orbiter is already set up the same way.

It may also allow some extra features that would not have been possible.

Anyway, over the next few weeks I will add and test it out. Hopefully it will work.

 

[Urwumpe]

I think it is a bit overkill for this small problem, but if the simple solutions can't solve it...its time for the big ones.

 

[tl8]

Good News:

I appear to have Bullet compiling and not crashing Orbiter. All that happens now is that a world is created and stepped through time.

Next step is to add and remove a rigid body (Harder than one expects as I have to scale down a bit). Once that is done I will add joints and link the visuals to the physics.

 

[tl8]

Update: Not much. I have been trying to animate the towbar and have not had much success. The code is now a big mess so I will need to clean it up soon.

My plan is to get this proof of concept working, then start on finishing the thing.

For those interested here is the latest dll and the latest code. You won't be able to compile the code due to the physics library.

Soon I hope to be asking for features....

 

[Grover]

it would be good to see this addon work perfectly.

ill put this DLL into my clean orbiter installation and give it a go.

thanks

 

[tl8]

Bullet is now linked with Orbiter!!!

Now just to get it working how it should and start a code clean up.

 

[tl8]

Small progress. Most of the bugs between bullet sim and the Orbiter sim are fixed (mainly getting the angles look right). A little luck on getting the right behaviour but nothing stable yet.