Ocado’s robots offer ‘safe pair of hands’ for packing shopping

Online retailer Ocado has developed a robotic system capable of packing bags with all the care and dexterity of a human.

UK-based online grocery retailer Ocado has for some years been at the forefront of artificial intelligence (AI) and robotics from a retail perspective, thanks to the work of its Ocado Technology unit.

As well as developing in-house robotic solutions for its customer fulfilment centers, the team is participating in two EU-funded Horizon 2020 projects, with a view to devising smarter warehouse robots: SoMa and SecondHands.

This week, Ocado has unveiled a new robotics solution that uses suction, computer vision and sensors to pack shopping items safely into bags.

Demand for smarter warehouse robots

Picking up and packing groceries is a routine task for humans, but it’s full of complicated decisions not easily replicated by an industrial robot. In a fast-moving fulfilment center handling an average of 260,000 orders per week, any automated system has to be able to think on its feet – so to speak.

Ocado deals with a huge range of products. In a high-speed game of retail Tetris, a successful packing robot has to take into account each item’s weight, shape and orientation when placing it among other objects.

The retailer’s latest solution is “conceptually simple“; its main features are a suction cup on the end of an articulated arm. But beneath the surface, there’s much more going on. The algorithm controlling the robot must have a grasp of more than just the items in front of it. It needs to understand the difference between the green crates that come from the storage department and the red ones that are designated for customer delivery, for starters.

It also needs to know and locate the “optimal grasp point” of every item within the crate. At the same time, it has to scan the delivery crate for free space and place multiple products with care, and often the right way up, too.

How 3D vision systems bridge the gap

Instead of taking on the monumental task of modelling every single item in the product catalogue, Ocado Technology developed a 3D vision system capable of identifying the best grasp points of each item. The suction cup is lowered down to the selected point and uses a vacuum cup to carefully pick it up and move it to the delivery crate.

While that is happening, the system confirms that the product is the correct one. It then works out the best orientation before placing it into the bag.

To make sure products aren’t damaged during the process, sensors built into the robotic arm consider the weight at hand.

In a blog post, Ocado Technology points out how much time and effort was saved by devising the 3D vision system. “The fact we found a way to bypass modelling our SKUs also meant that we could pick a greater range of items than many industrial picking systems. All in all, the system is streamlined and flexible, and our robotics team are very proud of the progress they have made so far.”