Resources / Core Concepts
Robot Arms in Warehousing: What They're Good At, and What Still Trips Them Up
Robot arms in warehousing excel at palletizing and station picking, while clutter, soft goods, and varied catalogs still trip them up. What to check first.
6 min read
· Updated
November 2026
At a glance
What it is
A robot arm is a programmable manipulator with three or more axes. In warehousing it works as a system with a gripper, cameras, and control software.
Where arms excel
Palletizing uniform cases, station picking, sortation induction, and packing or kitting where items arrive in a known place and orientation.
What trips them up
Varied catalogs no single gripper covers, cluttered bins, deformable items, clear or reflective packaging, and exceptions like double picks and empty slots.
A robot arm is a reprogrammable manipulator with three or more axes that grasps, moves, and places items, working with an end effector, cameras, and control software. Robot arms in warehousing palletize and depalletize cases, pick at fixed stations, induct items onto sorters, and pack or kit orders.
Robot arms have stacked cases onto pallets in distribution centers for decades. Ask an arm to pull one polybagged T-shirt from a crowded tote, drop it in a mailer, then reach for a glass bottle, and the job looks very different. The arm itself is rarely the weak point. The gripper, the cameras, and the software behind them decide whether the work gets done.
This explainer covers the arm types you'll find in warehouses, the jobs each does well, what still trips arms up, and how fixed and mobile arms compare. Use it to sort vendor claims about "robotic picking" into what the hardware can do and what still depends on your items.
What is a robot arm in warehousing?
The International Federation of Robotics, following ISO 8373:2021, defines an industrial robot as an automatically controlled, reprogrammable, multipurpose manipulator programmable in three or more axes, either fixed in place or fixed to a mobile platform, for use in automation applications in an industrial environment. In a warehouse, the arm is one part of a working system:
The arm: joints and links that set reach, payload, and range of motion.
The end effector: the gripper, suction cup, or tool that touches the item.
Perception: cameras and depth sensors that find items and check results.
Control software: grasp planning, motion planning, and the connection to your WMS.
Robot arm types you'll see in a warehouse
The IFR classifies industrial robots by mechanical structure. Four of its classes, plus collaborative arms, are the ones you're likely to see on a warehouse floor.
Articulated arms
An articulated robot has at least three rotary joints. Six-axis versions are common because they can approach an item from many angles, reaching into bins, around obstacles, and across a pallet. They are the workhorse for palletizing, depalletizing, each-picking cells, and arms mounted on mobile bases.
SCARA arms
A SCARA robot has two parallel rotary joints that give it compliance in a plane. It moves quickly and precisely across a flat work area with limited vertical travel. That suits fixed-station pick-and-place, but A3 notes its work envelope can be difficult to control, with restricted volume coverage.
Delta (parallel) robots
Delta robots hang above a conveyor, with several arms joined at a single tool. They are built for very fast, light pick-and-place, such as placing small items into trays or cartons on a packaging line. Payload and reach are small, and they need product presented on a belt below them.
Cartesian and gantry robots
Cartesian robots move along three linear axes, X, Y, and Z. Gantry versions span a work area from above. They are rigid and precise, which suits large, repetitive moves such as layer handling or tote transfers. The trade-off, as A3 puts it, is that they need a large volume of space to operate in.
Collaborative arms
"Collaborative" describes how an arm is used, not its shape. Many collaborative arms are articulated arms with sensing that limits contact force. The 2025 revision of ISO 10218 dropped the term "collaborative robot" in favor of collaborative application, because only the actual use can be confirmed as collaborative. Our guide to collaborative robots in the warehouse covers what that means for safety.
SKU fit check
Send an email to
What robot arms are good at in a warehouse
Palletizing uniform cases. Known case sizes and repeatable stacking patterns are close to ideal for an arm.
Depalletizing. Single-SKU pallets are well understood. Mixed pallets are harder: researchers working on supermarket depalletizing describe detecting, recognizing, and locating heterogeneous cases, which demands strong perceptive capabilities.
Station picking. When a goods-to-person system brings totes to a fixed arm, the arm picks into order containers without traveling.
Sortation induction. Placing parcels or apparel onto a sorter one at a time, with the label facing the scanner.
Packing and kitting at fixed stations. Placing items into boxes, pouches, or kit trays when presentation is consistent.
The pattern is simple. Arms do best when items arrive in a known place and orientation, and when the item set changes slowly.
What still trips robot arms up
Grasping across a varied catalog
No single gripper handles everything. In the first Amazon Picking Challenge, 8 of the 13 teams that scored used some form of suction and the winner used suction alone, but the published analysis notes that suction-only grippers have limited ability to manipulate items. Porous materials, mesh, and loose polybags don't seal well, and small or slippery items are hard to pinch.
Clutter and occlusion
Real bins aren't neatly arranged. Items overlap, hide in corners, and present their labels facing away. The arm has to choose a grasp that lifts exactly one item without disturbing the rest.
Deformable items
Apparel, pouches, and soft goods change shape when touched. A 2022 review in IEEE Robotics and Automation Magazine describes manipulating deformable objects as considerably more complex than rigid objects and still an open research problem, in part because a deformable item has far more degrees of freedom than a rigid one.
Hard-to-see packaging
Clear plastic, shrink wrap, and reflective or metallic packaging can confuse depth cameras. That affects both grasp planning and item verification.
Exceptions
Double picks, dropped items, damaged packaging, mislabeled products, and empty slots that inventory says are full all happen in production. An arm that can't detect and recover from them pushes the problem downstream to your team.
Fit call
Send an email to
The end effector matters as much as the arm
Suction: fast on flat, sealed surfaces such as cartons and rigid packaging.
Two-finger grippers: good for rigid items with clear edges, such as bottles and tubes.
Paddles and underneath support: for flat, soft, or awkward items that are hard to pinch or seal.
Tool changers: let one arm switch grippers between items or tasks.
When you evaluate an arm-based system, ask what share of your catalog each end effector covers and how quickly the arm switches between them.
Fixed arms vs arms on mobile bases
A fixed arm needs the work brought to it, usually by conveyor, goods-to-person storage, or a person. That works well for stable, high-volume flows, but it ties the arm to one task and often to new infrastructure. Mounting the arm on a mobile base lets it go to the work instead. For the trade-offs in depth, see our guide to mobile manipulation.
Site walkthrough
Send an email to
Questions to ask before you buy an arm-based system
Which of my SKUs has it picked, and what share of my order lines do they represent?
Which end effectors does it use, and how does it switch between them?
Does it need items presented in a fixed position, or can it pick from my existing bins?
How does it detect and recover from double picks, drops, and empty locations?
What infrastructure does it need: conveyors, guarding, fixtures, floor space?
Which safety standards and risk assessment apply to the full application?
Where Avatar fits
Avatar robots pick from traditional racks and bins and drop into your existing workflows, so the robot goes to the work instead of the work coming to it. See the Avatar picking page.
Frequently Asked Questions
What are robot arms used for in a warehouse?
Robot arms in warehousing handle palletizing and depalletizing, picking at fixed stations, inducting parcels and apparel onto sorters, and packing or kitting. They do best when items arrive in a known place and orientation. Arms mounted on mobile bases can also pick from shelving and bins across the building.
What is the difference between a SCARA and a six-axis robot arm?
A SCARA arm has two parallel rotary joints and works mainly in a flat plane, which makes it quick at simple pick-and-place with limited vertical reach. A six-axis articulated arm can approach items from many angles, so it handles bins, pallets, and awkward positions better.
Why do robot arms struggle with some items?
Most difficulty comes from grasping and perception rather than the arm. Polybags, mesh, and porous items don't seal for suction, soft goods change shape when touched, and clear or reflective packaging confuses depth cameras. Cluttered bins add occlusion on top of all three.
Is a collaborative robot arm safe to use next to people?
Not automatically. Under the 2025 edition of ISO 10218, collaboration is a property of the application, so the whole setup, including gripper, items, and layout, needs a risk assessment. Collaborative applications also tend to limit speed and payload so that any contact stays within safe force levels.
Sources
1. International Federation of Robotics (n.d.). Industrial robots: definition per ISO 8373:2021. IFR.
2. International Federation of Robotics (2016). World Robotics 2016, Section 1.2: Industrial robots, definition and classification. IFR.
3. Association for Advancing Automation (n.d.). SCARA vs. Cartesian Robots: Selecting the Right Type for Your Applications. A3 Industry Insights.
4. Association for Advancing Automation (Mar 20, 2025). Updated ISO 10218: Answers to Frequently Asked Questions (FAQs). A3.
5. Arpenti, P., et al. (2020). RGB-D Recognition and Localization of Cases for Robotic Depalletizing in Supermarkets. IEEE Robotics and Automation Letters, 5(4), 6233–6238. doi:10.1109/LRA.2020.3013936.
6. Correll, N., et al. (2016). Analysis and Observations from the First Amazon Picking Challenge. arXiv:1601.05484 (IEEE Transactions on Automation Science and Engineering).
7. Zhu, J., Cherubini, A., Dune, C., et al. (2022). Challenges and Outlook in Robotic Manipulation of Deformable Objects. IEEE Robotics and Automation Magazine, 29(3), 67–77.
Keep Reading
Core Concepts
Mobile Manipulation
Mobile manipulation puts a robot arm on an autonomous mobile base. Learn where it's used, how R15.08 classes it, and how it compares for warehouse picking.
3 min read
Read guide →
Core Concepts
Collaborative Robots
Collaborative robots in warehouses, explained: what ISO 10218:2025 changed, the four collaborative safety methods, and how to assess shared work.
5 min read
Read guide →
Picking
Piece Picking Automation
Piece picking automation explained: what a robot must do to pick one each, which grippers suit which items, and the checklist to complete before a pilot.
5 min read
Read guide →
Ready To Scale Your
Workforce?
Tell us your hardest jobs. We’ll show you what an Avatar robot can do on day one.
