\N

Robots struggle with the messy complexity of the real world, particularly with tasks like picking up objects from items jumbled together on a shelf.

But in only its second year, a contest to build a warehouse picking robot has seen a threefold improvement in the winning bot’s work rate.

The robot arms taking part in this year’s Amazon Picking Challenge performed far better than bots in the preceding competition.

Here you can see the winning bot made by Team Delft, a partnership between researchers from the TU Delft Robotics Institute and the company Delft Robotics in The Netherlands.

Team Delft came first in the picking contest, where bots scored points for correctly selecting individual items from those bundled together on shelves. The items represented a cross section of products found in Amazon’s warehouses, and ranged in shape and size from clothing to a toothbrush.

The team trained its bot to recognise the 40 Amazon warehouse items by feeding a deep learning system a database of 3D scans of the items. Doing the grasping was a robot arm with seven degrees of freedom — a term that reflects its ability to move in any direction in 3D space, as well as yaw, pitch and roll. The arm used high-quality 3D cameras for vision and a combined gripper/suction cup, designed in-house.