about CCD implementation and information

[Cak3Hol3]

Hi ^^,
im gathering informations about collision detection that prevents objects from tunneling.
I can't find any official paper work about CCD. I always thought that it was an algorithm, and it seems its only a name of a method to solve a problem.

So when im thinking about implementing collision detection with tunneling prevention, am i free to implement any collision detection algorithm like mikowski's sum or SAT or whatever and on that any TOImpact method that i pick to current needs? or is CCD described somewhere and i dont know something?

also what are the obstacles and pitfalls in (simple)implementing such system/algorithm?

also are there any good books about it except Gino's Van Den Bergen or Christer's Ericson?

Yes i want to make my own implementation.

sry for teh english q;

[Dirk Gregorius]

SAT is an algorithm to detect overlap between convex shapes and is often used to compute full contact manifolds when the shapes are in contact. GJK is an algorithm to compute the closest points between two disjoint convex shapes. GJK uses the Minkowski sum between the two shapes.

Conservative Advancement (CA) is often used to compute the TOI between two convex shapes. An explanation of the algorithm can be found in the PhD of Brian Mirtich. For pure linear motion see the paper by Gino about ray-casting. Implementations for linear and general convex cast can be found in Bullet. Box2D has a 2D implementation of CA.

[Oscar Civit Flores]

Hi,
You may also be interested in the work of Stephane Redon (and collaborators) on interval-arithmetic based continuous collision detection.

Also, a google-search reveals this article about bullet's CCD approach with references (at the end) to the some relevant CCD research:

http://www.continuousphysics.com/Bullet ... ection.pdf

Oscar