The age of applied robotics
Amazon has unveiled Amazon Sparrow, its new warehouse robot designed to replace packers, the men and women whose job is to put the items usually brought by the pickers — who themselves are increasingly being replaced by robots that bring the shelves containing them to the corresponding position on the periphery of the warehouse — into a packet with bubble wrap, close it, stick a label on and place it in a cart or on conveyor belt.
Sparrow makes it possible to automate the two most important functions in a warehouse: moving products and packing them for outbound distribution. And while the movement of products has long been carried out by robots that lift shelves and transport them, creating futuristic environments in which humans cannot enter because the aisles are just the right width for the transit of goods without wasting storage space, handling and packaging was, in principle, much more complex, because it requires adaptive capabilities to handle objects of a very wide variety of dimensions, materials, etc. with equally varied shapes.
How to design a robotic arm to perform a task that requires so many elements of what we would call fine human psychomotor skills? There’s only one answer: machine learning. Algorithms capable of recognizing an object from a huge catalog with the references of all the products in the warehouse, and depending on its characteristics…