Hi Open-TeleVision community,
URML (urml.dev) is a small, Apache-2.0 language for describing robot intent: it validates a request against a capability manifest and a safety envelope, then dispatches. Open-TeleVision is a leading immersive VR teleop system for bimanual and humanoid demonstration collection, and URML is interesting to it in two ways that don't compete with immersive control.
Nothing here asks the project to adopt, host, or maintain anything. This is a request for comment.
Two seams: (1) shared-autonomy handoff -- inside the immersive session the operator can issue a high-level command; URML validates it against the robot's capabilities + envelope and dispatches, with full teleop as the correction path. (2) typed-intent annotation -- a demonstration is labelled with the URML primitives it realizes, so an immersive bimanual demo carries validatable typed intent next to the video. URML already models bimanual manipulation (an arm selector + a bimanual primitive), so a two-arm demo maps cleanly.
Two real questions: (1) Is a validated shared-autonomy handoff interesting inside an immersive teleop session? (2) Is labelling demonstrations with typed URML intent (including bimanual) useful for the data collected -- and which is the cleaner first seam?
Full write-up: https://github.com/URML-MARS/URML/blob/main/docs/rfcs/0482-televison-outreach.md
Thanks for Open-TeleVision; an immersive bimanual teleop system is a great place to think about validated handoffs and typed demonstrations.
Ido Yahalomi (URML, greenvh@gmail.com)
AI-assisted prose, maintainer-reviewed before posting (see VIBE.md). Human-only correspondence available on request.
Hi Open-TeleVision community,
URML (urml.dev) is a small, Apache-2.0 language for describing robot intent: it validates a request against a capability manifest and a safety envelope, then dispatches. Open-TeleVision is a leading immersive VR teleop system for bimanual and humanoid demonstration collection, and URML is interesting to it in two ways that don't compete with immersive control.
Nothing here asks the project to adopt, host, or maintain anything. This is a request for comment.
Two seams: (1) shared-autonomy handoff -- inside the immersive session the operator can issue a high-level command; URML validates it against the robot's capabilities + envelope and dispatches, with full teleop as the correction path. (2) typed-intent annotation -- a demonstration is labelled with the URML primitives it realizes, so an immersive bimanual demo carries validatable typed intent next to the video. URML already models bimanual manipulation (an arm selector + a bimanual primitive), so a two-arm demo maps cleanly.
Two real questions: (1) Is a validated shared-autonomy handoff interesting inside an immersive teleop session? (2) Is labelling demonstrations with typed URML intent (including bimanual) useful for the data collected -- and which is the cleaner first seam?
Full write-up: https://github.com/URML-MARS/URML/blob/main/docs/rfcs/0482-televison-outreach.md
Thanks for Open-TeleVision; an immersive bimanual teleop system is a great place to think about validated handoffs and typed demonstrations.
Ido Yahalomi (URML, greenvh@gmail.com)
AI-assisted prose, maintainer-reviewed before posting (see VIBE.md). Human-only correspondence available on request.