[ODE] Nothing but problems with anything involving joints/wheels

Sponge realities at gmail.com
Fri Mar 31 08:31:14 MST 2006


I hope I'm replying correctly to this mailinglist:).

The mu/friction thing is one good idea! I should have come up with
that myself. I will try that this weekend, and hopefully it will work
out nicely :). It's ideas/suggestions like these which I am looking

Many thanks already... I'll let you know how it works out :-)

On 3/30/06, Jon Watte (ODE) <hplus-ode at mindcontrol.org> wrote:
> It sounds like your problem is actually with modeling, rather than with
> the simulation as such. I have no idea where to find good references for
> how to model various physical phenomenon -- it seems to be a black art
> that I've never found to be particularly challenging, but a lot of
> questions on the list relate to this area. Some kind of cook-book might
> be useful..
> You can move without wheels. The main question is what your contact
> joints do. You can model a cart as a simple box that's dragged on the
> ground if you want -- just set the contact joint parameters correctly.
> For a horse dragging a cart, you could model the cart as a simple box.
> Turn on mu and mu2 in your contacts for the cart. Set mu to something
> small (0.01?) and mu2 to something bigger (0.3?). Set fdir to forward in
> the box's direction for contacts generated at the rear end of the box,
> and to the direction from the contact to the center of the horse for
> contacts generated at the front of the box, and an interpolation between
> the two for contacts generated between the two.
> Now, applying a force to the box will make the box try to go forward,
> and turn towards the horse.
> Now, once you have the cart working with forces added to the cart, you
> can couple it to the horse. This is likely a universal joint, where the
> joint connection point is the hinge (I'm assuming perfectly rigid tack
> here -- unlikely in the real world :-). Don't change anything else in
> the cart. Now, model the horse using whatever method you prefer. I'd
> probably model it as another box, taller, and try to keep it upright
> with an amotor.
> For contact joints you create for the horse, do the similar kind of
> fdir/mu/mu2 model, except aim the front of the horse towards wherever
> the user is steering. To make the horse (and the cart) move, apply force
> forward on the horse body. Make sure the mu of the contacts is low
> enough that you can glide forward -- but when the user is not pressing
> forward, you may wish to set the mu higher (and the same front and
> sideways) so that the horse comes to a rest.
> Cheers,
>                        / h+
> Sponge wrote:
> > I have spent *weeks* to get objects connected to each other like a
> > 'car and trailer' (horse and wagon). I have read every single post
> > using the GMane search


More information about the ODE mailing list