Mettle and Metal: The DARPA Robotics Challenge

(U.S. Defense Department graphic illustration by Jessica L. Tozer/Released)

(U.S. Defense Department graphic illustration by Jessica L. Tozer/Released)

You guys, we need to talk.

There’s been a fair amount of oh, shall we say, embellishment on the real mission of the DARPA Robotics Challenge.  Some on a science-fiction level of paranoia.  Are we creating some kind of alpha-robotic race designed to dominate the planet?

The answer is no.  Just…no.  Everyone calm down.  The roborevolution is not upon us.  Believe it or not, these robots are here to help.  Like, actually help.

Especially in places where it’s too dangerous for humans to go.

“The purpose of the program is really to develop technology that can help make us much more robust to both natural and man-made disasters,” explains Gill Pratt, program manager for the DARPA Robotics Challenge.  “In particular, we’re looking at robotic technology that can allow us to mitigate the extent of a disaster during the first hours and days while the disaster is still unfolding.”

The final round of trials is approaching this very weekend; December 20th through 21st, at Homestead Raceway in Florida.  It’s the ultimate test of mettle and, well, metal.  The cause, however daunting, is a righteous one.

An example, and indeed what directly inspired this project, was the Fukushima earthquake and power plant disaster.  A lot of bad things happened quickly, and due to the extreme risk to human lives, they weren’t able to mitigate the circumstance fast enough.

“If human beings had been able to go into the reactor buildings during the first 24 hours and vent the built-up gas that was accumulating inside of the reactors, those explosions that occurred might have been prevented,” Gill says.  “The disaster might not have been as severe.”

That’s only one example, he points out.  We don’t know what the next disaster’s going to be, and so the technology we’re trying to develop is really to allow human beings and robots, working together, to have an effect on evolving disasters in environments that are too dangerous for human beings to go into by themselves.

RoboSimian, concept print.  Looks friendly, eh? ( Photo from the DARPA Robotics Challenge website/released)

RoboSimian, concept, print. Looks friendly, eh? ( Photo from the DARPA Robotics Challenge website/released)

That’s where the robotics challenge comes in.  The program with over 100 participating teams.  The first event, in June of this year, was a virtual robots challenge, or a simulation challenge.

From that event, as well as some design reviews, they narrowed the field to 17 participants.  Of those 17, there are 13 that DARPA is funding, but there are four others that are working on their own funding.  All these teams are working together with one goal in mind: to make the ultimate disaster-response robot.

In order to win, the robot must be able to accomplish three things (isn’t that always the way?):

One: the robots have to work in environments that were engineered for human beings.  Including environments that has been degraded by an evolving disaster.

Two: the robots have to be able to use human tools.  And by “tools,” I mean everything from a screwdriver to a fire truck.

“The reason for that is that there are usually many tools in abundance in the area where a disaster occurs,” Gill explains.  “Usually those tools are there for maintenance and construction.  Wouldn’t it be great if robots could reuse those tools to have an effect on the disaster?”

Three, DARPA wants these robots to be usable by personnel that are experts in handling disasters without them having to be experts in robotics.  They also want to cut down on the amount of training that’s required to use these machines by improving the human-to-robot interface.

The 17 teams in the running for the top robot spot vary across a wide spectrum, from small businesses to universities, software firms and hardware firms, and also some participants from the U.S. government as well.  Two teams in particular are from NASA.  In total, five different countries are being represented.

Thor. A Norse god name with a robot body.  ( Photo from the DARPA Robotics Challenge website/Released)

Thor. A Norse god name with a robot body. (Photo from DARPA Robotics Challenge website/Released)

With names like Chiron, Kaist, Valkerie, SCHAFT, THOR, and Robosimian, these robots are named to win.  They’re also designed like works of mobile art, with flexible arms and legs, anthropomorphic heads, and hardened metal bodies made of shiny material.  But this is more than just a beauty contest.

These robots are being pushed to the limits, because the limits are where they need to operate.

It’s very important, Gill points out, to distinguish between what real robotics technology is like, and what you’d see in science fiction and in film.

Right now, for the most part, most robots are either working on stationary bases in factories, doing very clearly defined repetitive tasks, or they are in laboratories in schools, where they are in very controlled environments.  If they’re in the outdoors, they’re typically run through something called “tele-operation,” where a human being is dictating every move that they’re doing every 10th of a second, or even faster.

What DARPA is trying to do is to advance that technology, and move things from tele-operation, to something known as “task-level autonomy“.

So, instead of telling a robot to move forward a 10th of an inch, or to move left a 10th of an inch, you can just tell it to open the door.  Far more convenient, if you ask me.

“That’s the level of supervision that we believe would be most effective for human beings and these kinds of disaster-response robots to interact with each other,” Gill says.

The reason for needing task-level autonomy is that, often, communications during a disaster are severely degraded, he says.  The channel of information between the human being that’s operating the robot, and the robot itself, is very poor.

“So, if you can’t say very much back and forth between the human being and the robot, you have to give the robot a little bit more smarts to be able to do these tasks a little bit on its own.  Of course, the robot still needs to be told what to do, but we’re just talking about things at the level of opening a door.”

Unfortunately, that’s not where the robotics field is right now.  Robots are roughly at the same level of mobility and dexterity as a one-year-old child.  So there’s a lot of falling.  A lot of dropping things.  They’re also quite slow.  In general, they need to try things many, many times to get them right.  That’s about where the field is now.

“What we’re doing with the DRC trials is we’re getting a calibration point,” Gill says.  “We’re trying to understand what the state of the art of the field is now.”

Chiron looks like serious business.  (Photo from the DARPA Robotics Challenge website/Released)

Chiron looks like serious business. (Photo from the DARPA Robotics Challenge website/Released)

So, the robots will be slow.  This is why they get half hour to do each of the eight tasks that are part of the DRC trials.  For example, a typical task is to go through three doors.  Imagine taking thirty minutes to walk through three doors in a row.  That’s the kind of time they are allowing for the machines to do.  So what you’re going to see is robots moving quite slowly.

But that’s because this is a baseline for what is to come.  The plan is to continue to improve, grow and program, so that the amount of human-to-robot work ratio leans far more on the robot side than it does now.

These robots will do a number of tasks, from driving a utility vehicle over a short course, turning and traveling over rough terrain.  They will have to negotiate obstacles, like removing rubble from a doorway, then walking through doors, climbing ladders.  Even picking up a tool and cutting through a WALL.  Each event gets scored, and the robot with the highest score wins.

Just in time for the holidays, too.

The results will be released the day after the trials, and I anxiously await those results.  Not only because of my pro-machine stance, but because of what this could mean for humanity as a whole.  Put ‘em to the test, DARPA.

Let’s see what these robots can do.

Follow the DRC Challenge on Twitter with #DARPADRC

Jessica L. Tozer is a blogger for DoDLive and Armed with Science.  She is an Army veteran and an avid science fiction fan, both of which contribute to her enthusiasm for technology in the military.

Follow Armed with Science on Facebook and Twitter!


Disclaimer: The appearance of hyperlinks does not constitute endorsement by the Department of Defense of this website or the information, products or services contained therein. For other than authorized activities such as military exchanges and Morale, Welfare and Recreation sites, the Department of Defense does not exercise any editorial control over the information you may find at these locations. Such links are provided consistent with the stated purpose of this DOD website.