We need to be able to split concave polygons into multiple convex sections in order to properly perform collision detection on them.