Added fixed models.

[S-Dafarra]

No description provided.

[traversaro]

Thanks!

@gabrielenava do you think this can conflict with icub-gazebo-wholebody ?

[traversaro]

That (no hands) is not ideal, but I see that you got it from the existing model.config.in, so if we remove it we will remove it from both.

[S-Dafarra]

I can also quickly add the models with the fixed feet

[S-Dafarra]

cc @nunoguedelha

[gabrielenava]

@gabrielenava do you think this can conflict with icub-gazebo-wholebody ?

I don't think so, because in master and devel branches of icub-gazebo-wholebody the fixed iCub models are pointing to iCubGazeboSim. But it seems there are old branches that go towards this direction, see https://github.com/robotology/icub-gazebo-wholebody/tree/feature/useModelsFromCAD.

I propose an optional alternative: maybe we can edit the already existing fixed models to point the @ROBOT_NAME@ model, to reduce the number of duplicated models.

[traversaro]

I can also quickly add the models with the fixed feet

Mhh, note that can be more tricky as if I recall correctly you need to hardcode the height of the waist in the fixed model sdf to correctly write the constraints on the feet.

[traversaro]

@gabrielenava do you think this can conflict with icub-gazebo-wholebody ?

I propose an optional alternative: maybe we can edit the already existing fixed models to point the @ROBOT_NAME@ model, to reduce the number of duplicated models.

To be honest, I am not a big fan of this alternative, as it would mean that to change the model that you use you need to reconfigure the project at the CMake level, and we are trying to support as much as possible also the use case of consuming this models from binaries.

[gabrielenava]

To be honest, I am not a big fan of this alternative, as it would mean that to change the model that you use you need to reconfigure the project at the CMake level, and we are trying to support as much as possible also the use case of consuming this models from binaries.

I think if we add the fixed models to icub-models we'll not have any conflict, the only not so nice effect will be that the number of models spread between the three repos (icub-gazebo, icub-gazebo-wholebody, icub-models) will increase. But at the Gazebo level, if all the required paths are installed through the robotology-superbuild, it is not an issue to handle the different models location.

So it is a go for me :-)

[S-Dafarra]

Mhh, note that can be more tricky as if I recall correctly you need to hardcode the height of the waist in the fixed model sdf to correctly write the constraints on the feet.

Ah right. Does it make sense for it to touch the ground anyway?

[traversaro]

Mhh, note that can be more tricky as if I recall correctly you need to hardcode the height of the waist in the fixed model sdf to correctly write the constraints on the feet.

Ah right. Does it make for it to touch the ground anyway?

Yes, it is done to make sure that the feet are fixed, but in a way that they are aligned with the fixed ground of Gazebo. We can also just fix them on mid air.

[S-Dafarra]

Yes, it is done to make sure that the feet are fixed, but in a way that they are aligned with the fixed ground of Gazebo. We can also just fix them on mid air.

What if I avoid specifying again the pose? Does it take the one of the original model? If so, it would be enough to add the fixed links 🤔

[S-Dafarra]

That (no hands) is not ideal, but I see that you got it from the existing model.config.in, so if we remove it we will remove it from both.

I moved most of the logic directly inside the CMakeLists so that the other confing.in file was not necessary. In this way, we would have to remove the (no hands) from a single place.

What if I avoid specifying again the pose? Does it take the one of the original model?

The answer is yes, it takes the one from the urdf. So I also added the "feet_fixed" models. I noted anyway that the iCubGazeboV2_5_plus starts a few mm above the ground. I did not attempt to fix this.

[S-Dafarra]

I think this is ready to be merged (even if I cannot).

[traversaro]

Perfect, thanks @S-Dafarra !