Research-collab proposal: URML cross-citation of OAK-D / DepthAI as candidate perception substrate

[idoco2003]

Hi Luxonis team,

Posting this on Discussions (rather than Issues) since the framing is research-collab + composition rather than a bug or feature request. I'm Ido Yahalomi, maintainer of URML, an Apache 2.0 specification for substrate-neutral robot intent. URML's Layer-2 primitive vocabulary sits one layer above ROS 2 / PX4 / vendor SDKs.

URML's value proposition for Luxonis: a documented cross-citation between URML's measure + wait_for primitives and DepthAI's perception streams. URML's measure(depth) + measure(orientation) + wait_for(object_detected) compose directly with OAK-D streams (stereo depth + IMU + on-device AI inference). The substrate-neutral story: a URML program written against OAK-D retargets to a different perception module by manifest swap. URML's value-add is the intent-validation + cross-substrate layer above DepthAI.

OAK-D is the de facto affordable 3D-perception module for home / educational / research robotics. The audience overlaps with URML's educational + research + home-assistance waves directly.

License alignment: MIT on depthai-ros + depthai-python composes cleanly with URML's Apache-2.0 reference/. The proposal is documented cross-citation + a hermetic DepthAI-conformance test fixture in URML's conformance/ suite; NOT a stand-alone Luxonis adapter (perception modules sit inside platform adapters, not alongside them).

Full RFC: https://github.com/URML-MARS/URML/blob/main/docs/rfcs/0105-luxonis-oak-d-outreach.md

This is proposal-only, part of URML's Move #8 outreach (home-assistance robotics + parts). Eight targets in this wave. URML has sent ~100 RFCs across eight outreach moves to date; single-digit substantive responses.

Feedback we'd value

1. Cross-citation appetite. Is Luxonis open to a documented cross-citation in URML's reference/ runtimes + conformance suite, naming OAK-D as a candidate perception substrate?
2. Primary-variant manifest. Which OAK-D variant (OAK-D S2, OAK-D Pro, OAK-4) is the right primary target for URML's documented mapping?
3. DepthAI v3 trajectory. Is DepthAI v3 the right substrate target, or should URML's cross-citation target v2 for stability?
4. Conformance lane. Open to a URML conformance line on depthai-ros or depthai-python README?
5. Educational + research profile co-design. URML's RFC-0011 / RFC-0012 raised the broader profile-design question; Luxonis's perspective from the affordable-perception side would inform the future Spec RFC.
6. GitHub Discussions vs Issues. I posted on Discussions; if Luxonis prefers Issues for this kind of RFC, happy to mirror.
7. Anything else.

Thanks for the open-source posture across depthai-python + depthai-ros + oak-hardware. URML's substrate-neutral story is much cleaner with an active, MIT-licensed perception substrate like OAK-D.

— Ido Yahalomi (URML maintainer, urml.dev)

---

Authoring disclosure: URML is the invention of Ido Yahalomi. The outreach prose is AI-assisted (Claude, under the maintainer's review). See VIBE.md. The maintainer reads and approves every post before it ships. Reviewers who prefer human-only correspondence are welcome to say so.