This robot wants to put MacGyver to shame

Georgia Tech researchers aim to build a "MacGyver robot" that knows how to use nearby objects to escape risky situations. Hopefully it won't sport a mullet.

The Golem Krang robot appears to be attempting to snatch Professor Mike Stilman's keys. Georgia Tech University

If MacGyver were trapped behind a jammed door in a burning room, he would use his shirt to filter the smoke, then craft an explosive from a paperclip and strand of hair to blow that baby open.

If today's most sophisticated robot found itself in the same conundrum, it would likely be unable to follow the famed secret agent's resourceful example. A team of Georgia Institute of Technology researchers hopes to change that.

They're working to equip machines to use objects in their path for high-level tasks, particularly those involved in tedious military operations. Robots are forging an increasing presence in military and civilian missions, with the U.S. military actively challenging roboticists to design robots for disaster relief.

"This project is challenging because there is a critical difference between moving objects out of the way and using objects to make a way," Mike Stilman, a Georgia Tech professor of robotics who's leading the research team, said in a statement. "Researchers in the robot motion planning field have traditionally used computerized vision systems to locate objects in a cluttered environment to plan collision-free paths, but these systems have not provided any information about the objects' functions."

To make robots more MacGyver-like (sans mullet, of course), the team plans to design a custom physics-based algorithm that lets bots identify an arbitrary object in a room, assess capabilities like its mechanical force properties, and turn it into a simple machine used to do things such as reach up high, brace a ladder against a bookshelf, stack boxes to climb over something, and build levers or bridges from random debris.

Stilman's team is collaborating with leaders from the field of machine learning. Once they've tested their system using computer simulations, the researchers plan to load their code onto Golem Krang, a humanoid robot designed and built in Stilman's lab for heavy service tasks. It can reach a human height of nearly 6 feet.

The team recently received a three-year, $900,000 grant from the Office of Naval Research for their work, but I'd submit that a MacGvyer housecleaning robot would make a lot of people awfully happy too.

Where there's smoke there'll be a MacGyver robot, Georgia Tech researchers hope. Georgia Tech University
 

Join the discussion

Conversation powered by Livefyre

Show Comments Hide Comments