I'm trying to program collision detection between moving convex polygons. To detect collisions, my program uses the GJK algorithm, and to resolve the collisions, it uses EPA. The problem with EPA is that it returns the vector to resolve the collision with the minimum distance, rather than returning the locations of the polygons at the time of the collision. Is there an algorithm that does that?

Edit: I realize my question is hard to understand, so let me rephrase/slightly change it: What continuous/predictive collision detection algorithm would you suggest?

  • 2
    \$\begingroup\$ It sounds like you're describing a problem commonly referred to as "continuous collision detection" — where rather than skipping the whole physics world forward one time step and checking/resolving overlaps in the final resulting state, we examine the full trail of space swept by each object in the course of its movement over the frame, identifying & resolving collisions by calculating the moment of contact with sub-frame precision. Am I understanding your question correctly? \$\endgroup\$ – DMGregory Dec 1 '15 at 4:58
  • \$\begingroup\$ @DMGregory yea that's pretty much what I mean. \$\endgroup\$ – EKW Dec 1 '15 at 18:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.