A VUB-Brubotics study tests whether AR headsets can help warehouse workers pick orders faster and with fewer errors, working side by side with a collaborative robot.
E-commerce keeps growing, and so do returns. A large share comes down to simple fulfilment errors: a missing item, the wrong quantity, the wrong product.
As part of the EU-funded OPENVERSE project, researchers at VUB-Brubotics built a complete AR-based order fulfilment system and tested it against the traditional paper picking list, in a setup where a collaborative robot brings and removes boxes.
The question was straightforward: do workers actually accept this technology, and does it help?
Twenty-seven participants completed the same picking task twice, once guided by an AR headset (Meta Quest 3) and once with a paper list, while a UR10e robot handled the boxes. The AR interface walked them through each item with names, pictures, quantities, and on-shelf location highlights, and managed safe communication around the robot's movements. Acceptance was measured using the Almere model, a established framework for human-robot interaction. What was found:
- AR was significantly more enjoyable and carried stronger social appeal than paper. Workers found it engaging rather than stressful.
- AR did not raise anxiety, even though it was new to every participant. Thoughtful interface design kept it intuitive.
- Paper still won on speed and on stated intention to use, mostly down to current headset comfort and the learning curve. Lighter hardware and familiarity are expected to close that gap.
Bottom line: AR shows real promise for reducing picking errors, but the hardware needs to mature before it replaces paper on the warehouse floor.
See the full task in action:
Burkiewicz et al. (2026),
Is augmented reality ready for the warehouse? A comparative study of user acceptance in human-robot collaborative order fulfillment,
Intelligent Service Robotics.
Read the paper