Garment-on-Hanger Handling in Distribution: Data for Hanging Goods

Why hanging goods are a separate case

Flat-packed apparel travels in cartons and totes and behaves like any other boxed good. Garments left on hangers are different: they hang from a hook, swing when touched, tangle with neighbors, and change shape when bunched together. The work involves hooks and rails more than shelves, and the grasp is often a hook or a bunch of hangers instead of a body of goods. That is why this page sits apart from garment folding and textile handling, which deals with folded flat goods, and from returns processing, where garments may be reprocessed.

The one source cited here, a review of warehouse order picking, describes picking systems generally and does not mention hanging garments. It is background about where labor sits in a warehouse. Task statements below are reasoning about how GOH operations are commonly arranged; confirm them with each supplier.

Task anatomy

1. Receiving. Garments arrive on hangers on rails, in trolleys or in transport containers designed for hanging goods; staff check them against documents.
2. Transfer. Hangers are moved from transport rails to storage rails by hand. Bunches of hangers are slid along, lifted or hooked over.
3. Storage and put-away. Garments are placed on rails by style, size or order group.
4. Order picking. Pickers pull specific garments from storage rails, hang them on a trolley or a conveyor rail, and push them to the next station.
5. Sortation. Garments are grouped by store or order, sometimes in a sorter that moves them along a rail.
6. Finishing. Items may be ticketed (see label application and relabeling), covered with a bag or a cover, and loaded.
7. Loading. Hanging garments go into a trailer or container with rails, or are moved onto a trolley; see trailer loading and load building for the boxed-goods version.

Pay attention to whether a supplier's operation includes hand transfers (the most manipulation-rich) or mainly automated rails with human supervision.

What footage and records can support

• Step and flow understanding. Segmented clips of transfer, pick and sortation support task structure.
• Perception of hangers and garments. Footage of rails, hooks and garments in realistic bunching helps detection of hanger types, garment categories and occlusion patterns.
• Order-to-garment linkage. Order and pick records can label which garment was taken for which order, usable as weak labels for identification.
• Attribute labels. Garment type, color, material and hanger style can be annotated; see apparel and textile attributes.
• Evaluation of perception. Held-out clips of tangled or crowded rails can test whether a system recognizes individual items.

Limits: what this data cannot support

Human video of hanging garments has no robot joint states, commanded actions or contact forces, and what matters in GOH work is how a worker feels and separates tangled hangers. Training a humanoid for this needs robot or teleoperation data, plausibly with tactile information; footage, SOPs or recognition data alone do not provide it. See what human demonstration video can and cannot do.

Specific gaps:

• Dynamic, swinging objects. These are deformable objects with added momentum. Motion depends on momentum and neighbors. A single camera at ordinary frame rates can blur or miss it.
• Occlusion. Densely hung garments hide each other, so labels per garment are hard to produce.
• Hook geometry. Hangers vary in hook shape, loops and clips; footage rarely records the type per item.
• Human reach. Workers use overhead reach and arm height that a humanoid platform may not match; see the human-to-humanoid embodiment gap.
• Seasonal change. Merchandise changes quickly, so appearance coverage ages.

Illustrative package: hanger transfer and pick clips

# Illustrative only - synthetic values, not an available dataset
package: goh_handling_v0
source: existing distribution-center video and pick records, 2 sites (site_G1, site_G2)
clips:
  - clip_id: goh_0291
    task: transfer_rail_to_rail      # receive_check | store | pick_to_trolley | sort | load
    garment_group: shirts_dresses_coats
    hanger_type: plastic_hook | wire | clip_hanger     # supplier categories
    density: sparse | bunched | tangled
    outcome: clean | dropped | snagged | re-hung
views: [aisle_fixed, close_hand_optional]
annotations: [step_segments, density_label, outcome_flag]
records_linked: order_line_hash_5521
excluded: brand marks (masked), worker identity, forces, robot actions

Counts in a real spec are the buyer's own choices. A pilot of "an arbitrary 150 clips per density level" would simply use arbitrary inputs, not minimums. Requirements depend on architecture, task, embodiment and intended use.

Quality and acceptance criteria

• Close views. Verify the hook-to-rail contact is visible in some clips.
• Density labels. Sparse, bunched and tangled states should each be present and labeled.
• Outcome labels. Drops, snags and re-hangs must be included, not edited out.
• Garment variety. Check coverage across coats, shirts, dresses and any heavier goods, as well as hanger types.
• Hold-out. Reserve whole sites or garment groups for evaluation; see training vs evaluation datasets.

Who may hold it

Apparel distributors, retail distribution centers with hanging-goods operations, and third-party logistics providers serving fashion clients may hold station video, order records and returns footage. See 3PL providers as warehouse data suppliers. Garment brands and retailers own the goods and may restrict recording of unreleased merchandise.

Permissions and provenance

Unreleased collections, brand marks, price tickets and customer order information are commercially sensitive. Ask who owns the goods, whether the brand has approved licensing, and whether order records contain customer data. Workers appear on video; blurring does not by itself settle permission questions. Evaluating a robot's ability to separate hangers is a different purpose from monitoring individual workers' pick rates, and a license should state the difference. See worker consent and identifiable people. This is not legal advice.

Questions to ask suppliers

1. Which GOH tasks are manual at your sites, and which are automated?
2. Can you show close views of hand-to-hanger contact?
3. What hanger types and garment groups appear?
4. Are tangles, drops and snags included, with outcomes?
5. Can clips be linked to order or pick records without customer data?
6. What camera types, frame rates and angles were used on the rails?
7. Who owns the merchandise, and is brand approval required to license footage?
8. How many sites and layouts are covered?

Discuss your requirements

If hanging goods are in your scope, describe the tasks, views 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. 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.

Describe the workplace demonstration data your robotics project needs

Sources

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