Egocentric Busbar and Limit Block Assembly Data

It is first-person video of one person placing a busbar and limit block, driving the screw sequence and removing the block, recorded from that person’s own point of view. A paid, consented contributor performs the task written in a skill spec, and each attempt is delivered as an episode with synchronized streams, hand pose and labeled action segments. It is human capture, not robot or teleoperation data, and not a fixed corpus you download.

The catalog row is written as a three-part sequence: "Place a busbar and limit block, drive the screw sequence, then remove the block." This page keeps to the row as written. It names no busbar, limit block, screw, driver or panel type, because the site names none, and treats anything more specific as a spec question settled in a Spec Pilot. It makes no claim about a trained model, benchmark or success rate.

One row counts. "Busbar + limit-block + screws" sits in the Precision Assembly Bench (electronics) environment, in the Manufacturing & Workshops sector, whose example line reads "Assembly bench: PCB/electronics, fasteners, fixtures, solder/crimp tools, small parts". The row is described as "Place a busbar and limit block, drive the screw sequence, then remove the block.", rated Hard, listed with "Multiview rig + handheld grippers" and Available on request.

Available on request is the site’s wording for a task scoped and recruited to spec when a buyer asks, as opposed to Active collection or Ready to deploy. So this page claims no delivered hours, stock of footage or published dataset for this task. The first table below reproduces the row, with neighbouring bench rows as context only.

The row says what is done, not what it is done with. It names no busbar, limit block, screw, driver or panel, gives no screw count, fastening order or tolerance, and states no torque value. Whether the sequence is a fixed pattern or varies between attempts is not stated. A buyer who needs a particular count, order, placement-only run or removal step as an accept criterion writes it into the skill spec, and a Spec Pilot proves it before you scale.

The Precision Assembly Bench lists 15 rows, all Available on request, each with a multiview rig and handheld grippers. The ratings are two Easy, ten Medium and three Hard, and the busbar row is one of the three Hard rows, alongside crimping a terminal and through-hole soldering.

"Screw-fasten housing" and "Torque to spec" are mentioned as context only. They are described as "Fasten a housing with a driver to torque." and "Torque bolts with a torque wrench.", both rated Medium and Available on request. They are single-purpose fastening rows; the busbar row is the one that combines placing, a screw sequence and removing a block. If your project needs several bench tasks, the precision assembly bench page treats all 15 rows; this page takes the busbar row on its own.

Fingers cover the small items they are placing, and the items cover the surface they go onto. The row adds ordering on top of that, because a sequence only makes sense when every step is visible and labeled. Three consequences follow, each already addressed on the site.

Occlusion: fingers disappear behind the items, which is why 3D hand pose carries a per-joint visibility flag. A hidden joint is marked as hidden rather than presented as a guess. Depth: the site states that small fasteners can fall below a depth sensor’s reliable floor, that hand pose stays valid at that range, and that metric depth on sub-centimetre parts is flagged rather than interpolated. No depth floor is stated for this row, so ask for it to be confirmed in the spec.

What video cannot show: the site lists no label for the outcome, and force, tactile and torque readings are not among the standard streams. The site also states that torque, force, tactile and fastener-type labels are not stated for fastening work.

The GR00T N1 paper says its human video datasets do not include explicit action labels, and that hand-held robot grippers speed up collection but leave a mismatch with the downstream robot. Firsthand captures people, not robots, so using the episodes for a specific robot means your own retargeting step.

For "Busbar + limit-block + screws" the catalog lists "Multiview rig + handheld grippers": a staged working set recorded by a multiview rig while the operator performs the task with handheld grippers. The catalog lists no stereo pair for this row, so this page does not describe the capture as stereo.

The wider stack the site documents is ten streams, hardware-triggered onto one PTP clock domain and aligned under 2 ms, with the rig calibrated at the start and end of every session. Which streams apply to a given busbar collection, how many cameras are used and what the grippers are is set in the skill spec; the site states no camera count, resolution or frame rate specific to this row.

Each attempt is delivered as one episode, and this row needs a decision up front about what one attempt is: the whole three-part sequence, or a single phase such as placement only. The site does not set that length, so the spec does.

Action segments divide the recording into spans carrying start_ns, end_ns, verb and noun, from a taxonomy of 97 verbs and 512 nouns, annotated to the frame and human-reviewed. A spec that says only "assemble the busbar" gives one long segment; one that names the phases gives a sequence you can slice, for example one span per screw driven if the spec defines it so. Other fields are hand_joints and hand_visible (21 joints per hand), contact_events, objects, and instance_masks at 848×480, plus int64 nanosecond timestamps, intrinsics and extrinsics per stream.

The site lists no label for the result: no seated-state, screw-order, defect or completion label. If your pipeline needs one, it is scoped in the spec rather than assumed. The site also states reject thresholds that apply to every collection: a cross-stream sync error above a 2.0 ms ceiling, the manipulated object out of frame for more than 15 percent of the episode, and more than 35 percent invalid depth pixels in the manipulation region. These are reject thresholds, not a stated pass rate.

Several pages touch fastening or bench work, but none takes the busbar row as its main topic, which is why this page exists. The comparison table below states each page as the site states it. For one fastening row against another, start with the screwdriving and fastening page; for the whole bench, start with the precision assembly bench page.

Only what the site already states applies. Every participant signs a written release before capture, and consent artefact IDs link each episode to it. Faces, plates and screens are detected and irreversibly blurred at ingest, speech is scrubbed of names, addresses and card numbers, and a human reviewer confirms no un-consented person is identifiable in any frame. Footage that cannot be de-identified is discarded, and the anonymization record is delivered with the data.

Printed text or markings on a panel or part are not a listed blur category, so this page does not claim they are blurred. Which bench, venue or region a program uses, and any safety practice, is not stated and is agreed in the spec. This is not legal advice, and the executed agreement controls.

The site states one published timeline from a signed spec to a first validated batch: a median of 48 hours where coverage is deep, and typically three to four weeks where it is thin. It does not say which applies to the Precision Assembly Bench, so ask when you request the quote. A tight spec is the quickest route to a fast, on-demand dataset.

The license is perpetual, irrevocable and worldwide, buyer-owned and non-exclusive by default, with exclusivity available as a separate term. Collections are billed per validated hour, rejected episodes and recaptures are never billed, and the site states no price list for this task.