Receiving and Inbound Checks

Task anatomy

Receiving is where goods enter a warehouse's records. A shipment arrives, is unloaded, and each pallet, case or parcel is identified and compared with what was announced. The announcement is an advance shipping notice (ASN), a record listing the expected units. A worker scans the unit's label, the system matches it to the notice, and the worker counts, inspects, sometimes weighs or measures, and then routes the goods to put-away, a quarantine area or an exception lane.

The task pairs two kinds of work: handling and verification. Handling includes pulling cartons from a pallet, tilting or opening them, and moving them to a staging position. Verification includes reading labels, matching codes, counting and judging condition. Robots that only handle can miss the check; systems that only read records can miss the physical state.

The failure modes are the interesting part: a label that will not scan, a pallet with mixed or unannounced items, a short or over shipment, crushed or wet cartons, wrong lot, and a unit received against the wrong notice. Exception rates differ by client and category, so collect them deliberately. This page sits in the warehouse cluster of the humanoid data guide. Unloading itself is covered in trailer and container unloading.

How identifiers tie the pieces together

The GS1 Serial Shipping Container Code is an 18-digit key that uniquely identifies a logistic unit such as a case, pallet or parcel, and is the key to ASN data. It is usually encoded in a GS1-128 barcode on the logistic label, or in EPC/RFID (GS1). Where a site uses it, a clip of a worker scanning a pallet can be linked to the notice line that the system matched, without storing the label's full contents. Without such a key, video and records are hard to align.

Order picking is the most labor-intensive and costly warehouse activity, and a review of the literature puts its cost at as much as 55% of total warehouse operating expense (de Koster et al., 2007). That review studies picking, not receiving, so it is cited only as context for the wider warehouse flow; it provides no evidence about receiving tasks or receiving errors. For picking, see each-picking demonstrations.

Which data helps which part

What the data can support

• Perception for inbound goods. Label location, damaged-carton recognition and pallet state.
• Verification logic. Language and rules from notices, mismatches and reason codes.
• Step understanding. Segmenting unload, scan, count, inspect and stage from video.
• Evaluation. Comparing a system's accept or reject decisions with recorded dispositions on held-out shipments.

What this data cannot support

• Robot actions or forces. Video and scan logs have no joint states, commands or contact data. Lifting a case differs for a human and a humanoid; see the embodiment gap.
• Client-confidential product data. Receiving records contain client item numbers, quantities and supplier names. They cannot be released without permission.
• Judgment criteria. Whether a dent is acceptable depends on the client's receiving standard, which may not be in the data.
• Rare events. Mixed pallets, hazardous leaks and severe damage are scarce and often handled off camera.
• Whole-site generality. A single dock's flow does not represent other dock layouts, systems or product mixes.

Illustrative package: receiving clips with ASN linkage

# Illustrative only - synthetic values, not an available dataset
package: receiving_checks_v0
site: dc_044
video: {station: dock_door_5, camera: fixed_wide_plus_handheld_scanner_cam, fps: 30}
records:
  asn_lines: asn_extract_000.csv      # item codes pseudonymized
  scan_events: scans_000.csv           # unit_key_hash, timestamp, result
  dispositions: disp_000.csv           # accepted, short, damaged, wrong_lot, hold
labels:
  steps: [unload, scan, match, count, inspect, stage]
  exception_type: [none, label_unreadable, short, over, damaged, mismatch]
  inspector_note: free_text_normalized
linkage: {unit_key_hash: salted_hash, clip_to_scan_offset_ms: measured}
excluded: [robot_actions, forces, client_item_descriptions]
rights: {licensor: operator_X, client_clearance: per_client}

Time alignment between camera and scanner clocks is a specific risk. Ask how offsets were measured and whether they are stored.

Sourcing and permissions

The usual holders are third-party logistics operators and distributors. Existing records are plentiful but footage often comes from security cameras with poor angles; commissioned recording lets you place cameras at the scanner and the count area and script exception cases. See 3PL providers as data suppliers and client-owned inventory rights.

Goods, notices and product data usually belong to the client rather than the warehouse operator. Ask which clients are in the footage and records, whether each has agreed, how worker consent was handled, and whether contracts bar use for model training. Redacting item codes or faces does not on its own settle those questions. This is not legal advice.

Quality and acceptance checks

Buyers set the thresholds. Checks worth writing into an acceptance test include the following.

• Match fidelity. For a sample of units, confirm that the recorded match result agrees with the ASN line and the video. Disagreements show clock offsets or manual overrides.
• Exception share. Tabulate exception types and check that rare ones are present in useful numbers. Handling of them is covered in warehouse exception handling.
• Scan context. Check that clips include the scan moment with the label visible, and whether barcode scanning workflows differ between handheld, fixed and RFID reads.
• Downstream link. If routing to storage matters to you, ask whether the record continues to the put-away location.
• Held-out design. Split by client, dock door or time so that evaluation shipments are never seen in training.

Questions to ask suppliers

1. Which receiving steps are filmed, and from which camera positions?
2. Can clips be linked to scan events and ASN lines through a unit key, and how are clock offsets recorded?
3. How many clients, product categories and carrier types appear?
4. What share of units are exceptions, and are the reason codes standardized?
5. Are damaged and mismatched units included, or only clean receipts?
6. Which fields in the ASN and receiving records are client-confidential, and what is the plan for pseudonymization?
7. Which clients have agreed to recording or reuse, and in what form?
8. What notice did workers receive, and who can license footage showing them?
9. What receiving standards (damage tolerance, count rules) apply to each client, and are they documented?

Discuss your requirements

If your project needs receiving data, describe the dock type, modalities and permitted uses to SourceX. SourceX looks for US businesses that 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, and nothing is contracted until a supplier agrees. The catalog entry for purchase orders describes a related record type.

Describe the workplace demonstration data your robotics project needs

Sources

• Serial Shipping Container Code (SSCC) — GS1 key identifying a logistic unit and linking to ASN data, the identifier receiving checks verify.
• Design and control of warehouse order picking: A literature review — Review describing order picking as the most labor-intensive and costly warehouse activity; picking-cost context only, it does not study receiving.