Robots will take over more of people’s jobs as they are designed to grab objects.
Currently, their lack of motor skills prevents them from performing certain objects, but that may soon change.
Researchers in California have found a way to speed up their artificial decision-making process 350 times.
Humans can manipulate objects with hardly any thought, but this seemingly simple task requires precisely coordinated movement and a complex understanding of an object’s size, pressure and fragility.
Robots can easily pick up objects for which they have been programmed, with something new that they cannot and can take time to master.
But Berkeley University students have cut a robot’s decision-making time from 29 seconds to 80 milliseconds, which could be a significant step forward for potential robot workers.
“This is an exciting new opportunity for robots to support human workers,” said Ken Goldberg, a professor of industrial engineering and senior author of the study.
“Warehouses are still mainly operated by humans, because it is still very difficult for robots to reliably grab many different objects.
“In an automotive assembly line, the same movement is repeated over and over so that it can be automated. But in a warehouse every order is different.
“Every second counts. Today’s systems spend up to half their cycle time on motion planning, so this method can significantly speed up the number of shots per hour.”
At the University of Singapore, researchers said in July that they had developed artificial skin that helps a robot detect what it is touching.
It can identify the shape, texture, and hardness of an object 1000 times faster than the human sensory nervous system, but they can take time to completely replace humans in warehouses.
Amazon has 100,000 robots that pack 600 to 700 boxes per hour, but they lack common sense.
Tye Brady, Amazon Robotics chief technologist, said in 2018 that when a pot hit the ground, the robots kept driving towards it, but couldn’t figure out what to do next.