Start a new topic

## Friction model

I'm using Vortex physics engine with VREP.

There are 7 types of the friction model in the material properties.

I want to know how the friction force calculated when the objects which are set different friction model are colliding.

The general rule that is applied to compute a VxContactMaterial from two VxMaterial is to take the values from the most compliant one (less stiff, less friction force etc). Indeed, what means "compliance" for a friction model. We have sorted the model from the most compliant to the less in the following order:

None

BoxProportionalLow

BoxProportionalHigh

ScaledBox

ScaledBoxFast

Box

Neutral

I set the three cases of the friction models.

1. ScaledBox - ScaledBox

2. ScaledBoxFast - ScaledBoxFast

3. ScaledBox - ScaledBoxFast

According to your explanation, No.1 and No.3 calculate the friction force as "ScaledBox" model.

But the simulation doesn't seem the same behavior.

What do you think?

My simulation is for a gripper which has micro spine.

First, for my understanding, I made a simple scene in VREP. I placed a cuboid on the 45 degrees inclined plane. Then I tested 3 types of simulations.

[1]ScaledBox - ScaledBox

[2]ScaledBoxFast - ScaledBoxFast

[3]ScaledBox - ScaledBoxFast

I understood your explanation like when objects which are set to different models collide, a friction force is calculated as less accuracy model.

So, I thought the same result come out with [1] and [3].

but, the displacements of the cuboid were different with [1] and [3].

Hi yusuke,

I made some tests to try to have an explanation of what you get and I found the following:

First, I trust in the rule I gave above regarding the model we chose when they are combined.

Second, a contact material has different properties depending on the direction of friction. This is seen in VxMaterial (see in the Material Table) under the sections "Linear Primary Axis" and "Linear Secondary Axis". Can you make sure that the test you made takes care of that. I say so because, I just realized during my investigation that, the default VxMaterial in the material table that comes with Vortex, doesn't have the same value along the two axes: one is scaledbox and the other one is scaledboxfast. You may have felt on this trap during your test.

On my side, I will take action to change this inconsistency in the material table for future release. On your side, please, verify if the test you made is done without such problem, and let me know if we can help more.

Best regards,

Michel

Thank you very much!