Re: [Algorithms] line-swept-sphere / AABB test

[Pierre T. <p.t...@wa...>]

Take the Minkowski sum of the AABB and a sphere with radius equal to the =
radius of the line-swept sphere. Then intersect this sum with the =
central line segment of the line-swept sphere. This is equivalent to =
intersecting the central line segment with an AABB-swept sphere with =
radius equal to the radius of the line-swept sphere. This is done by =
calculating the squared distance from the line segment to the AABB and =
comparing it to the squared radius. The distance calculation can become =
complicated; even calculating the distance between two line segments =
turns out to be difficult. Thus this approach is probably one of the =
least desirable.
=20
It is enough to say the LSS / AABB intersection test is similar to a =
segment / AABB distance query. Why introducing the (maybe confusing) =
Minkowski sum here ?

There's some code in MagicSoftware to do this. I'm using it in Opcode. =
Works well, and it's not that slow.

However a SAT approach might be better indeed. I think Charles Bloom has =
some code in Galaxy3 doing this (might need testing).

Pierre