Skip to Content
Artificial intelligence

Robotic Grocers Have Learned How to Handle Your Vegetables

The ability to gently grasp soft and unpredictable items will help pave the way toward fully automated grocery stores.
January 31, 2017

A robot that can pick up bags of fruit or individual vegetables may get us a step closer to fully automated grocery stores and warehouses.

Robots can put together cars just fine, but less ordered tasks still leave them stumbling. Recent advances have allowed robots to pick up hard objects in unstructured environments, but ask one to pass you a bag of bananas and things get difficult: the fruit varies in shape, it moves within the bag, and it’s delicate.

The online grocery retailer Ocado is trying to change that. It already operates several large and heavily automated grocery warehouses. But all of its goods are ultimately picked by humans, because robots can't handle the wide variation in size, shape, and structure among the 48,000 different products it sells.

Now, working with a number of universities, its robotics team has taught a soft gripper hand attached to the end of a robot arm to carefully pick up foodstuffs such as apples and bags of limes. Ocado's Alex Voica says that the device responds to changes in the shape of an item throughout the act of grasping, leading to “carefully choreographed movement of the hand in relation to the object.”

The system is also designed to identify specific contact points on the objects, then close around them like a human hand. The idea is to minimize the amount of bruising inflicted on fruits or vegetables.

Even so, the company has some problems to solve before it uses the devices throughout its warehouses. The experiments show that the device can pick up items placed alone on a flat surface. In a warehouse, on the other hand, they’d likely be piled up on top of each other in a crate, where they might shift as they were picked. Voica says that extra sensing systems and computer vision approaches could help the robot cope with those problems in the future.

There’s also the huge range of products to worry about. Fruit and vegetables are one thing, but a universal picking robot in an automated grocery warehouse would need to be able to grasp everything from, say, a bottle of wine to a croissant with ease and delicacy. Such a feat, says, Voica, “will be one of the hardest challenges to solve.”

Still, the company hopes to use the system commercially “in the near future.” With luck it won’t bruise any bananas.

(Read more: “Robot, Get the Fork Out of My Sink,” “The Robotic Grocery Store of the Future Is Here”)

Deep Dive

Artificial intelligence

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

Google DeepMind’s new generative model makes Super Mario–like games from scratch

Genie learns how to control games by watching hours and hours of video. It could help train next-gen robots too.

What’s next for generative video

OpenAI's Sora has raised the bar for AI moviemaking. Here are four things to bear in mind as we wrap our heads around what's coming.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.