What is the difference between GeometryObject's disableCollision and activeCollisionPair?

[joao-pm-santos96]

Hello all,

I am fiddling around with pinocchio's collision checking and got a question? Why are there two ways to disable the collision checking of a GeometryObject?

In here

pinocchio/include/pinocchio/algorithm/parallel/geometry.hpp

Lines 32 to 34 in ac0b1aa

	if(geom_data.activeCollisionPairs[cp_index]
	&& !( geom_model.geometryObjects[cp.first].disableCollision
	|| geom_model.geometryObjects[cp.second].disableCollision))

we can see that we can either disable the collision pair or set the disableCollision to False. But if I disable the collision of a given geometry obejct and then create the GeometryData associated with it, the collision pairs of that geometry remain active, which is a bit confusing... What am I missing?

Thank you!

[lmontaut]

Hi @joao-pm-santos96,
Depending on the state of your system, you might want to dynamically deactivate a certain collision pair (for that you can use geom_data.deactivateCollisionPair).
The disableCollision flag allows you to do that for all collision pairs involving a specific object, so you don't have to manually check in which collision pairs this object is involved.

Simple example: there is a wall in a scene; at some point in your simulation, the wall disappears (door opening, for example).