We started experimenting with picking up balls using the grab function and it works great, however, we only draw debug spheres for the hand positions. That means there is no real functionality in the hands except for the custom grab and throw we made.
We have seen several examples of hands knocking over cans or cubes etc. like this: Integration of Leap Motion in UDK (Unreal Development Kit) - YouTube
So we wonder how the general idea is behind this, is it some mesh attached to the hands, or the bones-class being used? There is not much documentation on these matters. Of course there is the possibility of setting a sphere (example) at the hands location which has collision embedded in it, but it’s a rather clunky and un-neccesary approach. Any help is appreciated.
You can do it either way.
It looks like they are getting input data (palm position, finger tips), and setting a static mesh to each of those positions. However, you could make a skeletal mesh and make the bone positions equate to your hand inputs.
For skeletal physics this may help : Physics Driven Animation in Unreal Engine | Unreal Engine 5.1 Documentation