WAM Add

[gaolongsen]

Hello, I add WAM 7 DOF robot in the package. The model test was without any problem. I look forward to being a contributor to the package!

[petercorke]

Hi @gaolongsen thanks for submitting the model, and I apologise for the very long time in responding.

All toolbox models are distributed via PyPI which has quite serious limits on the size of a package, rtb-data is already using "special permission" from the administrators due to is size, and I do need do be conscious of this.

I do not think that the SolidWorks models need to be included here, perhaps include a link to your own repo in a README.md file, that also references the two visualisations you include.

There are inertial parameters in this model, and I think they would be important to people who use a model of a robot like the WAM. Could you add that data?

I hope you can be a contributor to the toolbox.

[gaolongsen]

Hi Prof. Peter, I have made changes based on all your valuable suggestions. The new version keeps the same structure as other robot models that contains two folders: meshes + urdf. The inertia information for each link of the WAM robot with 7 DoF is included in the file "components.urdf.xacro"; also, for the Barrett hand, the inertia information for all links is in "hand.urdf.xacro". I tested the version in my local environment, and nothing went wrong. I'm looking forward to hearing your feedback and double-checking. Thanks for your time! It's my honor to contribute this pretty powerful tool library!

[gaolongsen]

@petercorke Hi Prof. Peter, would you mind taking the time to have a look at the PR? Thanks for your time!

[petercorke]

Thank you for your interest in RTB, and for putting together WAM support — a 7-DOF Barrett WAM model is a nice addition and we appreciate the contribution.

Unfortunately we're not able to merge this as-is, for a couple of reasons.

First, storage: the mesh/URDF assets this PR adds come to about 16.6 MB across 79 files, all of which would land in rtb-data — a separate PyPI package we're actively trying to keep lean rather than grow further, model by model.

Second, and independent of size: the WAM description here looks to originate from jhu-lcsr/barrett_model (Johns Hopkins LCSR's Barrett WAM/BHand URDFs), which is GPL-licensed (package.xml). Bundling GPL content directly into rtb-data — an MIT-licensed package we distribute on PyPI — isn't something we want to do without a closer look at the implications, even for data/mesh files rather than code.

Both concerns point the same direction: rather than vendor a copy into RTB, this is a good fit for robot_descriptions, the shared community package several of our newer models (UR5, UR10, Jaco, PR2, YuMi, ...) already load through at runtime. robot_descriptions explicitly accepts copyleft-licensed descriptions (its contribution guide only requires some open-source license, permissive or copyleft) because it works as a pointer to the source repository rather than a redistributed copy — it fetches from jhu-lcsr/barrett_model directly at install time rather than shipping the files itself, which is exactly the arrangement that avoids the licensing question above.

Adding it there is a small PR: register jhu-lcsr/barrett_model (at a pinned commit) in _repositories.py, add a small wam_description.py submodule, and list it in _descriptions.py's metadata. No mesh files travel through that PR at all. Once it's in robot_descriptions, adding the RTB-side WAM model class to load it is a small, welcome PR here — happy to help with that half if useful.

Separately, if a full mesh/URDF model is more than you need: we'd also very happily take a lightweight pure kinematic model instead — just the DH parameters or an ETS chain for the WAM, no geometry attached — as a small drop-in under models.DH or models.ETS. That has no storage cost at all and would be a welcome PR on its own, independent of anything above.

Closing for now given the storage and licensing constraints, but please do come back — either with a robot_descriptions entry for the full model, or a lightweight DH/ETS-only one, we'd be glad to add WAM to RTB's model catalog either way.