Science and Tech

A robot learns to find hidden objects

[Img #66523]

It’s pretty easy for us humans to organize the process of looking for a lost purse that’s buried under a pile of objects: we simply remove things from the pile until we find the purse. But for a robot, this task involves complex reasoning about the heap and the objects in it, which is very challenging.

Researchers at the Massachusetts Institute of Technology (MIT) in the United States previously demonstrated a robotic arm that relies on visual information and radio frequency (RF) signals to find hidden objects tagged with RFID tags (which reflect signals sent by an antenna). .

Building on that work, they have now developed a new system that can efficiently find any object buried in a lot of stuff. As long as some objects in the heap have RFID tags, the object does not need to be tagged for the system to locate it.

Algorithms in the new system, called FuseBot, reason about the most likely location and orientation of objects under the heap. FuseBot then finds the most efficient way to remove the obstructing objects and extract the desired object. In tests, this strategy allowed FuseBot to find more hidden objects than a conventional state-of-the-art robotic system, in half the time.

Members of the research and development team. From left to right: Fadel Adib, Tara Boroushaki, Nazish Naeem and Laura Dodds. (Photo: James Day, MIT Media Lab. CC BY-NC-ND 3.0)

This great speed in locating objects could be especially useful in an e-commerce warehouse. With the FuseBot system, a returns-processing robot could more efficiently find items in an unsorted pile, argues Fadel Adib of the research and development team. (Font: NCYT by Amazings)

Source link