Eaches picking with Brightpick mobile manipulators at Brightpick

Engineering & simulationPhotoneo 3D cameras / Photoneo vision technologyautomate 95% of the entire process

The Brightpick mobile manipulator uses onboard Photoneo 3D cameras to image totes pulled from shelves, enabling the robot to identify and pick specific items from random assortments without human intervention.

“they already know and have a solution for 3D vision and and robot guidance. So it's equipped with a photo nail camera. And so they pull the bin into the imaging area. They select an item out of the, uh, or more out of the bin for specific order…”

Engineering & simulationmachine learning picking enginearound 80% of items from the from the basket we see

Brightpick developed its own machine learning-based picking engine that allows the robot to generalize to unseen items, achieving high pick rates in unstructured grocery environments where items vary in appearance.

“Over the time, we develop our own mission learning, uh, a picking engine. Today we can tell that it's, you know, general enough to pick invisible, very high number of items. For example, in grocery ambient environment, it's around 80% of items from the from the basket we see.”

Engineering & simulationmachine learning / AInot possible 5 to 6 years ago

The company leverages recent advances in artificial intelligence and machine learning to generalize to unseen items, a capability they state was not possible 5-6 years ago and is critical for handling the variability of grocery items.

“all the things which are happening over the last 2 years, simply there, that was not possible 5 to 6 years ago, you know, it's like, this is a really serious breakthrough that simply your machine learning can generalize to unseen items, period.”