[ODE] fake car physics

[Jaroslav Sinecky]

This shouldn't be problem... just adjust the contact parameters between your
chassis geom (or maybe also wheels geom) and the wall giving it lower
friction
... or a cheaper way using force-dependent-slip (dContactSlip1,2). This
works quite ok for me setting the slip to 0.3 (this number can be specific
to my case though).

Hope this helps,
Jaroslav


> -----Original Message-----
> From: ode-bounces at q12.org [mailto:ode-bounces at q12.org]On Behalf
> Of David Salz
> Sent: Sunday, August 07, 2005 8:24 PM
> To: ode at q12.org
> Subject: RE: [ODE] fake car physics
>
>
> Thanks for the answer! I already know Carworld and it certainly
> helped me a
> lot. One major problem I have at the moment is that when I hit a
> wall my car
> bounces back, flips over or comes to a complete stop. While this may be
> correct, it is not what I want. I would like the car to keep
> "sliding" along
> the wall in the right direction - or maybe bounce back a bit, slow down a
> bit but keep going in the right direction anyway. Has anyone done
> something
> like this?
>
> David
>
> > -----Original Message-----
> > From: Megan Fox [mailto:shalinor at gmail.com]
> > Sent: Sunday, August 07, 2005 8:04 PM
> > To: David Salz
> > Cc: ode at q12.org
> > Subject: Re: [ODE] fake car physics
> >
> > (Most of these concepts can be found in "Carworld," an ODE demo that
> > was made to answer exactly this sort of question:
> > http://www.mindcontrol.org/~hplus/carworld.html)
> >
> > If you want something like Unreal Tournament 2004's or GTA's vehicles,
> > all you need is a box with 4 rays being cast downard at each of the
> > four corners.  You simulate a spring system at each of the 4 rays,
> > simulate the friction of the "wheels", force application to the body
> > due to said frictions, and everything else - since you're in control
> > of everything, you can determine exactly how it behaves.
> >
> > Ironically, it ends up being a lot more work to simulate a "fake" car
> > than a physically accurate car... but you kind of need to do most of
> > this work to make a nice-handling car anyways, and it's faster than a
> > full sim car.
> >
> > Concepts of note include:
> >
> > - anti-sway torque: as the body leans, you apply counter-torque to fix
> > the lean.  You can effectively disable roll-overs this way (just apply
> > unlimited counter-torque past a certain tilt point) if that's your
> > goal, and by controlling the torque application curve you can simulate
> > any number of easy or hard-to-rollover vehicles.
> >
> > - center of mass: use dGeomTransform's to make your center of mass
> > somewhere low.  If you want a weeble-people car, you can displace it a
> > good deal below the ground even.
> >
> > - turn stability: turning at high speeds can often cause rollovers in
> > a sim.  This is realistic, but unfun.  You can modify your max
> > correctional torque by the wheel turn to help eliminate this, but
> > still allow average "I just got T-boned" rollovers.  Incidentally,
> > this gives you GTA-esque "Steer into the lean to prevent just about
> > anything from rolling over, no matter how far it's tipped" physics,
> > which I find quite agreeable.
> >
> > -Megan Fox
>
> _______________________________________________
> ODE mailing list
> ODE at q12.org
> http://q12.org/mailman/listinfo/ode