# edge edge collision response

This topic is 3003 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

## Recommended Posts

Hey, guys, now I am doing some cloth simulation and I need to do some collision detection and reaction.

vertex - triangle is no problem, and also I can detect the potential collision of edge - edge, however, I don't figure out how to deal with the collision response problem, especially the edge - edge collision. I detect there will be a collision between two edges, and I can find the collision point, then what should I do?

If it is vertex - triangle, that will be easy, just calculate a new vertex velocity.

Thank you very much.

##### Share on other sites
I think perhaps the triangle-vertex case is not as simple as you think. You can't just recalculate the velocity of the vertex, because then the triangle won't respond at all to the collision. I don't know the precise solution, but perhaps carefully thinking through the triangle-vertex case will help you with the edge-edge case.

##### Share on other sites
thank you, and actually I know it's not that easy; I read some papers and you are right, we should consider the velocity of the triangle, but I just don't think that will be a "big" problem, since the response could be directly added to the vertex and triangle; however, in the edge - edge collision case, even though we can find some collision point, still we need get a idea how the response could added to the vertices of edges, instead of the collision point.

##### Share on other sites
collision response in general is not that easy. And certainly it's not a linear algorithm, except you're using verlet-intergration, but this is far from being good.
Any other method (LCP, jacobi-iteration on impulses, SI, etc...) is more complex then just implementing a single formula and applying it to a single contact-point.
You'll have to dig much deeper into the matter, if you really want to implement it. If you're just planning on having some kind of "physics", and you don't care a lot about precision, then go with verlet-intergration. It's simple to implement, but as already said, not very robust.
good luck

1. 1
2. 2
Rutin
19
3. 3
khawk
15
4. 4
5. 5
A4L
13

• 13
• 26
• 10
• 11
• 44
• ### Forum Statistics

• Total Topics
633744
• Total Posts
3013653
×