[ODE] Collisions via Minkowski sums

[Sergio Valverde]

Your exposition remind me the GJK algorithm for locating the closests
points between two convex sets. There is also a paper of SOLID's author
(don't remember the name, Gino I think) about finding the penetration
depth. Anyway, those algorithms are very prone to numerical innacuracies
so they require careful programming.

-----Original Message-----
From: Thomas Harte [mailto:thomasharte@lycos.co.uk]
Sent: jueves, 27 de marzo de 2003 15:43
To: ode-list
Subject: [ODE] Collisions via Minkowski sums


The built in collision stuff in ODE isn't really suitable for my purposes,
so 
I have been rolling my own. A little reading on the subject from various 
papers across the internet has revealed that forming the Minkowski sum 
of one of my objects and the negation of the other allows me to easily 
(given that my objects are broken into convex quantities) extract, in 
ODE terms, penetration depth and contact normal.

I haven't read anything on faking a contact 'point' as ODE likes to think, 
so at the minute my plan is to calculate the overlap region - which 
intuition tells me will also be convex - and take the centroid of that.

Has anyone on this list any experience in this sort of field? Am I thinking 
along the right lines?

-Thomas

When words aren't enough - Vodafone live! A new world of colour, sounds,
picture messages and information on your mobile. <a
href="http://ad.doubleclick.net/clk;4909903;7724245;q?http://www.vodafone.co
.uk/live">
Click here</a> to find out more.