Real-Time Physics Simulation Forum

I would thought it would be the natural extension to GJK for performing swept tests, I don't think it was anything new really. And I got my vector naming convention wrong :oops: I will try that in a test app to see if it works. I have an app already in place. And btw, the algo is wrong, The closest ...

I have toyed around the idea, what always gets me is how big to make the basis :? Similarly, if you use a triangle - tetrahedron, how to define it's shape and size? This will obviously affect the moment of inertia perceived. I like the solution, but it has shortcomings and is a bit fuzzy round the e...

All that sphere collision talk got me thinking a little. I am thinking of an algorithm to collide spheres against over convex shapes, which is similar to the GJK algorythm, and is an iterative method. However it requires square roots to find the point of contact (due to the sphere's nature). It's us...

code (untested) : // // solve vector equation : (P + V * t)^2 = r2 // equation in form : a*t^2 + b * t + c = 0 // a = (V . V) // b = (P . V) * 2 // c = (P . P) - r2 // returns first positive root. // // special cases : // ----------------- // if (c < 0), we have intersection -> returns 0 time. // if...

but yeah, that doesn't really help you with the problem. sorry!

car simulators is a hot topic, are really hard. Even to the level of trackmania. Trackmania is even harder since it's so extreme, they have to cheat (downforce, grip) to make the car controllable and the game enjoyable. Another good example would be Rollcage for achievable physics at very high speed...

no worries.

If you have sphere triangle tests working, there should be a sphere segment implementation somewhere (unless they use the 'ellipsoid collision' which is broken in any case). ok then... Sphere - vertex : sphere [O, r, D] // O : centre, r : radius, D, displacement vertex [V] // V : vertex position doi...

Thanks for the response and the link. If this is Renault, thank you very much for the pollycolly demo, I have been using it as reference for my project. It's really cool that the physics community is so active in the forums. Yeah, the problem is, that demo is old and there is a better method. :) . ...

If you need more details (implementation if you want), shout. I've got loads of code examples.

Yeah. Sphere-Box. You need to find which faces are candidates for collision. This should be relatively straight forward with a bit of partitioning. There can only be a maximum of three faces (say, the sphere is facing a corner) Then you simply test the box against each potential faces (same as spher...

You can extract overlap information from the SAT. as you say, the calculations are not normalised, but you can work with squared distances really easy. Then you take the smallest overlap of all for your separation axis. As for finding the contact points, http://www.gamedev.net/community/forums/topic...