[Enhancement]: Proposal: InterbotixAdapter for URML's substrate-neutral robot-intent language (X-Series, ROS 2 + legacy ROS 1)

[idoco2003]

Describe the solution you'd like

Proposing an InterbotixAdapter under URML's reference/cobot-runtime/ targeting interbotix_ros_manipulators plus the sibling interbotix_ros_core and interbotix_ros_toolboxes. The adapter routes URML Layer-2 primitives (move_to, grasp, release, measure, wait_for, report) and the industrial-profile extensions (pick_from, place_at, swap_tool) onto the X-Series joint-trajectory action goals and gripper services across ROS 2 Humble, ROS 2 Rolling, and (optionally) ROS 1 Noetic for the legacy lane.

URML is an Apache 2.0 specification for substrate-neutral robot intent at urml.dev. Interbotix sits where URML's educational, research, and US-federal-compliant lanes converge: the X-Series is de-facto research-arm hardware in US curricula (Stanford, CMU, MIT, Berkeley) and powers Mobile ALOHA's four-arm bimanual rig via four VX300S units. URML's value-add is the substrate-neutral vocabulary that lets a program written against an Interbotix arm retarget to a UR3, a Franka, or a WLKATA Haro380 without source changes, with static validation as the safety boundary.

This is proposal-only, posted as part of URML's Move #3 outreach (the affordable / desktop / educational tier). No adapter code in this PR. The per-arm manifest split across the X-Series (PX100 / PX150 / RX150 / RX200 / WX200 / WX250 / WX250S / VX250 / VX300 / VX300S), the gripper-variant schema, and the ROS 1 Noetic legacy-lane coverage are observable choices worth your input before shipping.

This RFC also cross-links to two of URML's Move #2 outreach RFCs: RFC-0040 (Hugging Face LeRobot) covers the policy library distributed in part on Interbotix hardware, and RFC-0056 (Stanford ALOHA) covers the Mobile ALOHA recording pipeline that runs on Interbotix arms. The two cross-links are observed in the RFC; no joint action proposed.

Full RFC with proposed package layout, per-arm manifest sketches, drawbacks, and alternatives:

https://github.com/URML-MARS/URML/blob/main/docs/rfcs/0064-trossen-interbotix-outreach.md

Feedback we'd value

1. Manifest granularity. One manifest per X-Series arm, or a single parametric interbotix_x_series manifest with a model: field?
2. ROS 1 Noetic legacy lane. Should URML's adapter cover ROS 1 Noetic, or target ROS 2 only?
3. Gripper-variant surface. Is the gripper: field's single-value design sufficient, or do you recommend a richer schema for finger-position vs. custom end-effectors?
4. Adapter home. URML repo (reference/cobot-runtime/), Interbotix org as a contributed example, both?
5. Bimanual coordination. Mobile ALOHA coordinates four VX300S arms via ALOHA's recording stack. Path for a coordinate(arm0, arm1, ...) Layer-2 primitive at the ROS 2 layer, or stays at the recording / policy layer?
6. Conformance lane. Open to a URML conformance line on the package README or in the X-Series documentation?

Thanks for the Interbotix ROS 2 packages and for the cross-ROS-version build matrix. The US-domiciled provenance plus the academic-curriculum reach is exactly what URML's open-standard story needs at the research tier of Move #3.

— Ido Yahalomi (URML maintainer, urml.dev)

Platform

No response

Is your feature request related to a problem? Please describe.

No response

Suggested code

No response

Additional Info

No response