Sunday, November 29, 2009

A new hope

So I finally did what I should have done a long time ago - I started coding physics again. All those ideas that were lingering in the back of my head in the Meqon and Ageia days but I never had time to try will finally materialize, be implemented and posted on this very blog, along with pictures, videos, demos and general ideas around physics engines.

So far I've focused on rigid body dynamics, and more specifically the contact manifold and solver. I'm a big fan of GJK, so there's no surprise it makes up the backbone of the collision detection engine.

The solver is of sequential impulse type with warmstarting (just like everybody else), but has a big bag of tricks that accumulated over the years. I put a lot of effort in friction this time, since I really wanted a realistic friction that does not creep, but actually sticks and comes to rest, even in non-trivial poses.

Allright, here's the obligatory box pyramid:



No comments:

Post a Comment