For Ground Handling Operators

Air Cargo Automation for Ground Handlers

DUND automates the hardest part of the cargo floor: breaking down aviation ULD pallets and rebuilding them into client-ready ground pallets. Computer vision maps the pallet, robot arms do the lifting, and every box is read and matched to the manifest — running 24/7 under operator supervision at up to 200 boxes per hour and an average of 2 ULDs per hour per system, in real airport cargo environments including the Schiphol region in the Netherlands. Multiple systems can be installed in a single facility, scaling throughput linearly with every added system.

200/h
Boxes per hour (per system)
2/h
Average ULDs / hour (per system)
5,000kg
Weight handled per pallet
0
Manual lifts in breakdown
The Problem

What automation actually solves on the cargo floor

01

Shift coverage & labour shortage

Night and weekend shifts are the hardest to staff, yet freighters land around the clock. An automated system runs the same speed at 03:00 as it does at midday, so the schedule stops depending on who shows up.

02

Manual lifting injuries

Boxes of 60+ kg, wet cartons and deformed cargo are handled by hand hundreds of times per ULD. Automating the lift removes the single biggest source of back injuries and sick days on the cargo floor.

03

Breakdown errors and re-sorting

Every box has to reach the right consignee pallet. Vision-based label reading matched against the manifest removes mis-sorts before the pallet leaves the build-up area.

04

No visibility into the breakdown

Manual breakdown produces no data. An automated system logs every box, weight and destination as it is handled, so throughput and discrepancies are measurable per ULD.

How It Works

From ULD arrival to transport-ready pallet

  1. 01

    Scan the ULD

    A multi-sensor vision array maps the incoming aviation pallet in 3D and plans the unloading sequence before the first pick — including mixed and irregular box sizes on the same pallet.

  2. 02

    Robotic breakdown

    An industrial robot arm with a vacuum gripper lifts each box — deformed, wet or tilted up to ±45° — onto the conveyor. No manual lifting is required at any point of the breakdown.

  3. 03

    Scan & match

    Labels are read automatically and matched against the manifest, so each box is assigned to the correct consignee before it moves on.

  4. 04

    Build client pallets

    A palletizing robot builds dedicated ground pallets per client in parallel — transport-ready when the last box lands.

Operating Envelope

What the platform handles

DUND automated ULD depalletizing — operating specifications
Throughput (per system)Up to 200 boxes / hour · avg. 2 ULDs / hour
Weight handled per palletUp to 5,000 kg
ScalabilityMultiple systems per facility — throughput scales linearly
ULD size handledUp to 3,200 × 2,600 × 3,000 mm
Boxes per pallet100–200
Max box weight60+ kg
Cargo conditionDeformed, wet or tilted up to ±45°
Operation24/7, operator supervised

The figures above are per system: one system delivers up to 200 boxes per hour and handles up to 5,000 kg per pallet. A facility can run multiple systems in parallel, scaling throughput linearly with every added system. Each system is designed around real aviation cargo, not idealised cartons: mixed box sizes on one pallet, damaged and moisture-affected packaging, and perishable freight such as flowers where handling time directly affects product value.

Questions from the Cargo Floor

Frequently asked questions

Which ULD types and sizes can the system handle?

The system handles aviation pallets and containers up to 3,200 × 2,600 × 3,000 mm, with a total load of up to 5,000 kg per pallet. A single ULD typically holds 100–200 boxes, and mixed box sizes on the same pallet are handled without reconfiguring the system.

What happens with damaged, wet or badly stacked cargo?

The vision system plans each pick from a 3D map of the actual pallet, so deformed cartons, moisture-affected packaging and boxes tilted up to ±45° are still picked. When cargo exceeds the defined grasp envelope, the system enters a controlled pause and escalates to the supervising operator rather than executing an out-of-tolerance pick.

How fast is it compared to a manual breakdown team?

One system processes up to 200 boxes per hour, averaging about 2 ULDs per hour, at the same pace on a night shift as at midday. Because the figures are per system, a facility can install multiple systems in parallel and scale throughput linearly.

How many operators are needed?

The breakdown itself requires no manual lifting. An operator supervises from a control panel and intervenes only for exceptions, so staffing shifts from physical handling to supervision.

How does the system know which client pallet a box belongs to?

Labels are read automatically during handling and matched against the manifest, so each box is assigned to the correct consignee before it moves to the build-up area. A portable scanning table complements the system: place a box on it and in about a second it shows which client pallet it belongs to.

Does it integrate with our existing manifest and warehouse systems?

DUND develops the vision system, grasp planning, robot software and operator interface in-house, so the integration is adapted to your manifest data, ULD mix and warehouse layout rather than forcing the operation around an off-the-shelf robot.

What data does the system produce?

Every box, its weight and its destination are logged as it is handled, so throughput, handling time and discrepancies become measurable per ULD — data that a manual breakdown does not produce at all.

Where is the system deployed?

Deployments run in real airport cargo environments, including the Schiphol region in the Netherlands, on live freight rather than idealised test cartons.

Deployment

Built by one team, from code to cargo floor

DUND develops the vision system, the grasp planning, the robot software and the operator interface in-house. That means the system is adapted to your warehouse layout, your ULD mix and your manifest data — instead of forcing your operation around an off-the-shelf robot. Operators supervise from a control panel; the system asks for help only when cargo falls outside its envelope.

Deployments run in real airport cargo environments, including the Schiphol region in the Netherlands. A portable scanning table complements the system: place a box on it and in about a second it reads the labels, matches them against the manifest and shows which client pallet it belongs to.