🔒 Locked

collision theory 2d

Started by Zeophlite Dec 27, 2004 at 4:49 AM 3 replies 1.7k views

Original Post

Dec 27, 2004 at 4:49 AM

Fig A - http://vicious.sixlabs.org/figA.bmp Fig B - http://vicious.sixlabs.org/figB.bmp In fig A, the circle moves ONTO the red rectangle, and the collision is detected However, in fig B the circle moves OVER the red rectangle, and the collision is missed It has been suggested to limit the distance the circle moves, so it can't skip collisions, however, I don't want to do that What other forms of collisions are there any other possible ways to detect collisions? Really appreciate a link, Thanks in advance

Nik02

Dec 27, 2004 at 5:22 AM

One could use a swept-circle check there. That is, form a "2d capsule" from the collider start and end points, and check the capsule's - instead of only the start and end circles - collision against the other collision targets. By start and end, I mean the collider positions in adjacent timesteps of the simulation.

Hint: A capsule can easily be modeled parametrically by a line and a thickness [wink]

Zeophlite

Dec 27, 2004 at 5:37 AM

I see, is there a web site describing that?
How fast is it?

jetro

Dec 27, 2004 at 9:00 AM

Check out this: http://www.gamasutra.com/features/19991018/Gomez_1.htm

Nik02

Dec 27, 2004 at 12:25 PM

The corresponding 2d techniques can easily be derived from the 3d ones - just remove the last element of the vectors involved [smile]

Topic Locked

This topic has been locked by a moderator. New replies are not allowed.