Each year, Amazon gathers together researchers at a Robotics Challenge where machines must pick and stow objects. It’s a tough job, but one that could ultimately help the e-tailer to fully automate its warehouses. This year the task was made even harder than usual—teams had only 30 minutes to familiarize themselves with the objects, and had to pick them out of a jumble where items sometimes sat on top of others—to simulate warehouse conditions more realistically.
The winner, a robot called Cartman, was built by the Australian Centre for Robotic Vision. The device was one of the least human-like devices of all, with grippers moving in 3-D like a funfair claw crane. According to Dr Anton Milan, one of its creators, the device's computer visions systems were crucial to the victory. "One feature of our system was that it worked off a very small amount of hand annotated training data," he explained to TechAU. "We only needed just seven images of each unseen item for us to be able to detect them."