This picture is from one of Dirk Gregorius' presentations: enter image description here

The original diagram has only one contact point (the red one). The r_1 and r_2 vectors are required to produce rotation while solving the contact constraints. I agree that the r_2 vector is correct, but I think that the r_1 vector should point towards the blue point (which I've added myself)

It's not difficult to retrieve the blue point (just add the minimum translation vector to the red point), but I've seen contrasting depictions of what points the contact manifold consists of.

Excluding the parallel edge-edge case, should the contact manifold consist of two points (the deepest points of each shape) or just one?

  • \$\begingroup\$ When you simulate your physics using these two points as you think it should be done, what kind of results do you get? Does the simulation do what you want? Or does it produce a specific negative behaviour we can help you solve? When you test with both one-point and two-point manifolds for this case, does one strategy or the other give better results in the context of the rest of your simulation code? \$\endgroup\$ – DMGregory Sep 3 '20 at 11:19

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.