[ODE] Interacting with moving world geometry.

[Konstantin Voloshin]

Hello,

It's not hard at all as long as only contact joint is conserned. ODE-contact-joint
already support this functionality partially. I mean moving-surface parameters.
If you add possibility to supply surface moving speed along 3-rd axis, you are
ready for one-way interaction with moving geometry. That is you create
contact-joint with 0 for "static" body, and with proper surface-movement
parameters.

However, for complete support for static-moving bodies there's a need to
special branches in all joints. Probably, a nice API for this would be to have
all this done automatically for you hidden inside ODE internals, but
personally I'd like much better a convenient way (possibly supported by API,
as for user-defined geom-types) to add user-defined joint-types to ODE with
a set of handy functions for common evaluations, performed inside joints.

Konstantin


----- Original Message ----- 
From: "Megan Fox" <shalinor at circustent.us>
To: "Alen Ladavac" <alenl-ml at croteam.com>; <ode at q12.org>
Sent: Tuesday, May 18, 2004 12:45 AM
Subject: RE: [ODE] Interacting with moving world geometry.


To resurrect an old thread - has anyone implemented anything like the
kinematic body concept that Alen describes (in an open-source can-share-it
way, that is)?  Essentially, some sort of analog to Tokamak's AnimatedBody.

-Megan Fox

-----Original Message-----
From: ode-bounces at q12.org [mailto:ode-bounces at q12.org]On Behalf Of Alen
Ladavac
Sent: Friday, April 16, 2004 11:04 AM
To: ode at q12.org
Subject: Re: [ODE] Interacting with moving world geometry.


The clean way to do it would be to add concept of kinematic bodies to ODE -
those that do not accept force and torque, but can still have velocities. If
you do it as moving static geometries, it turns out other bodies bounce of
wrong, because velocity is not incorporated into the joint equation. If, on
the other hand, you put in a very high mass, you either don't get real
kinematic behavior (if the mass is not high enough), or get a singular
inverse mass matrix (if mass is too high).

The change shoud not be very hard to do, mostly a few changes in the contact
joint so the velocities get calculated even though the joint force cannot
influence it. This is similar to the problem recently mentioned with a
stand-up object. Just that here you need all 6DOFs to be kinematic.

HTH,
Alen

_______________________________________________
ODE mailing list
ODE at q12.org
http://q12.org/mailman/listinfo/ode