In order for household robots to be truly useful, it would be great if they could go and get items for you, without having to be shown where those things are. Thanks to research being carried out at Georgia Tech, that may someday be the case. A robot there is now able to search out hidden objects – as long as they've been labelled first.
Ordinarily, robots identify items using technologies such as computer vision systems. These only work as long as the object is out in the open, however, plus they can be confused by cluttered environments.
UPGRADE TO NEW ATLAS PLUS
More than 1,200 New Atlas Plus subscribers directly support our journalism, and get access to our premium ad-free site and email newsletter. Join them for just US$19 a year.UPGRADE
Instead, the Georgia Tech team applied ultra-high frequency radio-frequency identification (UHF RFID) adhesive tags to items such as a medicine bottle, TV remote, phone and hair brush. They additionally equipped a PR2 robot with a specialized search algorithm, that utilized two articulated antennas mounted on its shoulders.
As the robot moved about a room, trying to locate each of the items, its antennas would swivel to assess the strength of that object's RFID signal – the stronger the signal, the closer the object must be. By incrementally moving towards the source of the signal based on those readings, the robot was able to locate all of the items.
It is hoped that once perfected, the technology could find use in applications such as personal care robots, which have to select and deliver the correct medication to their user.
"This could allow a robot to search for, grasp and deliver the right medication to the right person at the right time," said former Georgia Tech student Travis Deyle, who developed the system with Prof. Matthew Reynolds and associate professor Charlie Kemp. "RFID provides precise identification, so the risk of delivering the wrong medication is dramatically reduced."
The PR2 can be seen performing its search, in the following video.
Source: Georgia Tech