Robots Learning Better Ways to Ask Clueless Humans for Help

If robots need help with something, they can now just ask you

2 min read
Robots Learning Better Ways to Ask Clueless Humans for Help

A lot of the time, robots seem pretty dumb to us humans. It's not entirely a surprise, then, that a lot of the time, humans apparently seem pretty dumb to robots. If you're a robot, it turns out to be surprisingly difficult to get a human to assist you with tasks, so researchers at MIT are teaching robots to politely ask for very specific kinds of help.

You'll recognize these KUKA youBots from our ICRA coverage earlier this year: they can autonomously assemble (at least one specific piece) of IKEA furniture. But being youBots, they're relatively limited in what they can reach, so unless the furniture bits and pieces are set up just the way they need them to be, they're going to need help from a human. It's not always obvious to the human what kind of help the robots needs, however, which is why these robots are using an inverse semantic algorithm to generate human-friendly requests for help:

The robots are able to detect failures by themselves, determine exactly what sort of help they need, and then translate those requests into something understandable and actionable by a human. Since it's algorithmic and not pre-programmed, the 'bots should be able to ask for help under conditions that they haven't already experienced.

This is obviously of tremendous importance to robotic IKEA furniture assembly, but if we can look beyond that (just for a second), there's a huge amount of potential here. Robots are great at completing most parts of most tasks, but inevitably there are one or two steps in whatever you want a robot to do where it's significantly more likely to fail. Giving robots the ability to recognize these failures and then intelligently ask for assistance could open up many more tasks to at least partial automation, and it's likely to have the most impact in variable, unstructured environments. You know, like your house.

A paper accompanying this video (which we don't have a copy of yet, unfortunately) has been authored by Stefanie Tellex, Ross A. Knepper, Adrian Li, Thomas M. Howard, Daniela Rus, and Nicholas Roy.

[ MIT CSAIL ]

The Conversation (0)

How the U.S. Army Is Turning Robots Into Team Players

Engineers battle the limits of deep learning for battlefield bots

11 min read
Robot with threads near a fallen branch

RoMan, the Army Research Laboratory's robotic manipulator, considers the best way to grasp and move a tree branch at the Adelphi Laboratory Center, in Maryland.

Evan Ackerman
LightGreen

“I should probably not be standing this close," I think to myself, as the robot slowly approaches a large tree branch on the floor in front of me. It's not the size of the branch that makes me nervous—it's that the robot is operating autonomously, and that while I know what it's supposed to do, I'm not entirely sure what it will do. If everything works the way the roboticists at the U.S. Army Research Laboratory (ARL) in Adelphi, Md., expect, the robot will identify the branch, grasp it, and drag it out of the way. These folks know what they're doing, but I've spent enough time around robots that I take a small step backwards anyway.

This article is part of our special report on AI, “The Great AI Reckoning.”

The robot, named RoMan, for Robotic Manipulator, is about the size of a large lawn mower, with a tracked base that helps it handle most kinds of terrain. At the front, it has a squat torso equipped with cameras and depth sensors, as well as a pair of arms that were harvested from a prototype disaster-response robot originally developed at NASA's Jet Propulsion Laboratory for a DARPA robotics competition. RoMan's job today is roadway clearing, a multistep task that ARL wants the robot to complete as autonomously as possible. Instead of instructing the robot to grasp specific objects in specific ways and move them to specific places, the operators tell RoMan to "go clear a path." It's then up to the robot to make all the decisions necessary to achieve that objective.

Keep Reading ↓ Show less