[ODE] Generic Surface Contact Properties

Matthew Harmon matt at ev-interactive.com
Thu May 19 09:14:36 MST 2005


Hi all: 

I’m just starting to play with ODE and I’m trying to organize my thoughts
regarding surface contact properties and was wondering if anybody had 2c to
share about this issue. Here’s where I am: 

In a simple world I’d be able do define different surface types with values
for mu, slip1, slip2, etc. Then, when two objects collide, I’d look up their
surface data and combine them based on some algorithm, producing a
convincing collision. Of course, that's not how the world works.  These
values must be derived empirically on a pair-by-pair basis. In other words,
there isn’t a simple equation to factor two surface "mu" values and come up
with a good resulting mu for the contact. Or is there? 

- Has anybody had luck generating contact parameters by somehow combining
values from two different surface descriptions? Something like
mu=(muA*muB)/(muA+muB). I know an approach like this isn’t "correct", but
perhaps something is close enough to produce a good "feel".

- Has anybody taken the approach of creating a big table containing
hand-tuned values for the interaction of any two different surface types?

- For erp, cfm and bounce values I was thinking it may be good to just take
the larger of the two values and use that.  i.e. a contact is always as
squishy as it's squishiest object.

Any other thoughts would be most appreciated.






More information about the ODE mailing list