Air Cargo Build-Up and Breakdown: Data for Contoured Loads

Why air cargo differs from palletizing

Mixed-case palletizing in a distribution center stacks uniform cases on a flat pallet. Air cargo build-up places very different pieces (crates, drums, boxed machinery, bagged goods, odd-shaped parcels) onto an air pallet or into a unit load device (ULD) whose outline is not a simple box. The handler often has to fit freight inside a limited contour, balance the load, place heavier pieces low, and work within the time before a flight cut-off. Breakdown reverses this: the load is taken apart, checked against paperwork, and routed on.

This page cites no source specific to air cargo. The order-picking literature review cited here does not discuss it, so treat that paper as general warehouse background only. Any statement about contour limits, restraint, weight distribution or regulations belongs with the supplier's own written procedures. This page draws no conclusions about aviation or safety rules; it gives you questions to ask, and compares the task with trailer loading and load building and case picking and mixed palletizing.

Task anatomy to look for

1. Freight acceptance and staging. Pieces arrive, are checked against documents, and are staged by flight or destination.
2. Piece measurement and weighing. The handler may need dimensions and weight for planning; see dimensioning and weighing records.
3. Build plan. A planner or experienced builder decides what goes where, perhaps with loading software, perhaps by eye.
4. Placement. Pieces are lifted or moved onto the base, often with powered conveyors, lifts or hand equipment. Heavy and awkward pieces are placed first, and gaps are filled with smaller ones.
5. Securing. Nets, straps or covers hold the load; the specifics vary by unit and are not asserted here.
6. Documentation and release. The unit is tagged, weighed again and moved to its position. Discrepancies go to an exception route; see warehouse exception handling.

Breakdown involves the same elements in reverse (compare trailer and container unloading), plus identifying each piece and finding damage.

What data can support

Limits: what this data cannot support

None of these assets contains robot joint states, commanded actions or contact forces. Training a humanoid to build or break down a load requires robot or teleoperation data, and the heavy, awkward items involved are a poor fit for evaluating on real hardware without care. Footage, SOPs or recognition data alone do not provide control training. See human demonstrations vs teleoperated robot demonstrations.

Task-specific concerns:

• Equipment does much of the work. Lifts, conveyors and handling aids may carry the load, so video shows human supervision of equipment more than hand manipulation. Check what share of clips shows human hands on freight.
• Time pressure and safety rules shape behavior. Under a deadline, workers skip or compress steps; recordings reflect real operations, not ideal ones.
• Contoured fitting is spatial reasoning. Whether a piece fits is partly a planning problem. A video of the result does not reveal the rejected alternatives.
• Restricted spaces. Cameras in secure areas may be unavailable or limited, and recordings from them may be restricted.
• Heavy pieces are out of scope for many humanoids. What a humanoid can lift is a platform question; the footage cannot answer it.

Illustrative record: one build-up job

# Illustrative only - synthetic values, not an available dataset
build_job:
  job_id: abu_0178
  unit_type: air_pallet_A           # supplier's own unit categories
  pieces: "arbitrary example, 22"
  piece_mix: [crate, drum, boxed_machine, bagged_goods, parcel]
  plan_source: handler_planner_v1
  video_views: [overhead_fixed, side_fixed]
  events:
    - t: 00:03:12
      action: place
      piece_ref: piece_07
      note: "heavy piece low, center"
    - t: 00:41:08
      action: reposition
      reason: "contour clearance"
  outcome: accepted | re-built | rejected_at_check
  damage_log: none | scuff | crush
excluded: shipper names, consignee data, robot actions, forces

The numbers above are synthetic. Any sizing arithmetic drawn from them, such as a count of jobs, uses arbitrary inputs, not minimums; requirements depend on architecture, task and embodiment.

Quality and acceptance criteria

• Plan-to-outcome linkage. Each video should be tied to a build plan and a result, so you can label plan adherence and rework.
• Piece identification. Check whether individual pieces can be tracked from staging to placement.
• Coverage of difficult loads. Include crates, drums and odd shapes, not only boxes.
• Hold-out design. Reserve entire facilities or unit types for evaluation; see training vs evaluation datasets.
• Redaction effect. Test whether masking shipper marks removes cues you need.

Who may hold it

Ground handlers, cargo terminal operators and freight forwarders may hold station video, load sheets and exception logs. Airlines may hold documentation of their own operations. Supply for this task is narrow and may require commissioned recording or specialist partners; see existing assets, new collection or specialist partners. Access to secure areas and shipper consent are likely to be the gating factors; this page states no security requirements, so treat both as diligence items for the facility operator.

Permissions and provenance

Freight belongs to shippers and may be sensitive, high-value or restricted; markings, consignee names and manifests can identify them. Ask who can license footage and records, what shipper permissions apply, and whether the facility restricts recording. Workers appear on video. Masking or blurring does not by itself settle permission questions. Evaluating a robot's perception of load contours is separate from monitoring individual workers' speed. See worker consent and identifiable people. This is not legal advice.

Questions to ask suppliers

1. Which unit types and freight mixes do your recordings cover?
2. How much of each recording shows hands on freight versus equipment operation?
3. Are build plans and load sheets available, and in what format?
4. Do any cameras show inside the stack or the contour edges during fitting?
5. Are rebuilds, rejections and damage incidents included?
6. What restrictions apply to recording or exporting from secure areas?
7. How are shipper and consignee identifiers removed, and what does that remove from the images?
8. Who can license the data, and are shipper permissions needed?
9. Were recordings made for training, security or claims, and under what notice?
10. Who approves a capture team's access to restricted or secure areas, and what escort, screening or credential steps would the facility require?

Discuss your requirements

For a task this specialized, begin by describing the unit types, modalities and permitted uses to SourceX. SourceX looks for US businesses that may hold suitable material, checks the data and the supplier's licensing permissions, and coordinates the license and delivery. If no existing asset fits, SourceX can scope a commissioned collection of new recordings. A request does not guarantee a match; SourceX follows up about potential sources. Start from the humanoid training data guide or the warehouse hub, and see supply chain and logistics records.

Explore sourcing and licensing options with SourceX

Sources

• Design and control of warehouse order picking: A literature review — general warehouse background; it does not cover air cargo.