At the 2014, International Conference on Intelligence Robot and Systems, Tech PhD graduate Dr. Travis Deyle, Professor Charlie Kemp and University of Washington Professor Matthew Reynolds presented a new Radio Frequency Identification (RFID)based search algorithm for locating objects in a 3D environment with the help of robots.
The research was implemented on a PR2 robot and it can locate objects correctly with the help of RFID tags.
“In robotics, one of the hardest challenges is to know the identity of objects,” said Feyle. “It is easy for a person to walk into a room and locate stuff, but in robotics it is one of the toughest questions out there.”
An RFID tag on an object can give the robot the information it needs to identify and locate the object in a room.
Long range RFID, implemented by Deyle and his team, can be used to great effect in a room.
“It works when objects are out of sight also,” Feyle said. “If an object is inside a cabinet or a behind the sofa, the robot can locate it using long range RFID. Also, there can be multiple objects in the room from 100 to a 1000 and the robot will be able to locate the precise object.
“Using RFID and tags we can only locate a rough area of where the object is located in the environment,” Deyle saud. “For precise object location we will have to use some different technology like cameras or laser. The tag could be used to determine its precise location.”
While working on the project, Deyle and his team implemented radar techniques from the 1950s for localizing and pin point searching.
“We looked up stuff in 1950s radar literature which holds up even today and if we had better hardware we would’ve been able to implement more stuff from that literature for even better localization,” Deyle sajd. “A lot of people think RF is a black box but it really isn’t. There are simple things which builds this foundation”
One of the possible application for advances like these is medication delivery and for robotics in home care.