On Dec. 21, at a racetrack complex near Miami, eight robots took their first steps into the future, one that scientists at the Defense Advanced Research Projects Agency envisioned for them as the world watched the Fukushima Daiichi nuclear disaster unfold in 2011.
Soon after, they established the DARPA Robotics Challenge. The challenge’s first event took place in June with 26 teams from eight countries and whose last event will begin at the end of 2014 at the DRC Finals with a winning team, a robot, and a $2 million award.
DARPA launched the challenge to help develop robots that can work with people to respond to natural and other kinds of disasters.
The competition — consisting of robot systems and software teams — was designed to be very difficult, at a time in robot evolution when real robots generally work on stationary bases in factories doing clearly defined repetitive tasks or in controlled laboratory environments, said DRC program manager Dr. Gill Pratt.
Participating teams, down from 100 when the program began, to 16 at the DRC Trials 2013 held Dec. 20-21 at the Homestead-Miami Speedway in Florida, represent some of the world’s most advanced robotics research and development organizations.
The teams, DARPA officials said, are collaborating and innovating on a very short timeline to develop hardware, software, sensors and human-machine control interfaces that will allow their robots to complete a series of tasks chosen for their relevance to disaster response.
In Florida, while thousands of spectators watched and often cheered, the 16 international teams, representing industry, academia and NASA, guided their robots through as many as eight individual, physical tasks that tested the robots’ mobility, manipulation, dexterity, perception and operator-control mechanisms.
Eight teams scored the highest number of points for all tasks completed and are headed for the 2014 finals. Pratt said DARPA has $8 million to divide among the teams for further development, depending on individual team contract negotiations, to prepare for the finals.
The qualifying teams were:
1. 27 points, Team SCHAFT, Robot S‐One, lead organization: SCHAFT Inc.;
2. 20 points, Team IHMC Robotics, Robot Atlas‐Ian, lead organization: Florida Institute for Human & Machine Cognition;
3. 18 points, Team Tartan Rescue, Robot CHIMP, lead organization: Carnegie Mellon University, National Robotics Engineering Center;
4. 16 points, Team MIT, Robot Atlas-Helios, lead organization: Massachusetts Institute of Technology, Computer Science and Artificial Intelligence Laboratory;
5. 14 points, Team RoboSimian, Robot RoboSimian, lead organization: NASA Jet Propulsion Laboratory;
6. 11 points, Team TRACLabs, Robot Atlas, lead organization: TRACLabs Inc.;
7. 11 points, Team: WPI Robotics Engineering C Squad (WRECS), Robot Atlas-Warner, lead organization: Worcester Polytechnic Institute; and
8. 9 points, Team Trooper, Robot Atlas, lead organization: Lockheed Martin Advanced Technology Labs.
On the first day of the trials, speaking from victory lane next to the race track, DARPA Director Arati Prabhakar thanked the teams for their hard work, ingenuity and collaboration with DARPA.
“We will show the world what’s possible, what we need to work on, what it’s really going to take for robots to step up when disaster strikes,” Prabhakar said, adding that the teams would help build a better future with robotic capabilities.
Jyuji Hewitt, deputy director of the U.S. Army Research, Development and Engineering Command, called the teams’ efforts a reality check into technology’s “art of the possible.”
Dr. Brad Tousley, director of DARPA’s Tactical Technology Office, said DARPA was happy with the teams, the robots and the enthusiastic crowds.
“There are a lot of teams getting a lot of points,” he said during a press briefing after the first day of the two-day event. “From our standpoint that shows real progress.”
After the closing ceremony, during which Pratt awarded certificates to the eight winning teams, the scientist briefed the press, summing up the DRC’s results.
“The technology has shown itself to be more advanced than I had expected — not by much but by a little bit. The robots were more reliable than we initially expected — not by a lot but by a little. But that’s only half of the picture,” he explained.
“We’re not quite sure of the [date for the 2014 Finals] yet but we will advance the robots significantly, much more than we have up until now,” Pratt said. “We’ll see how far we get.”
In the Miami trials the robots had to complete as many as eight separate challenges.
These included driving a utility vehicle over a short course, getting out of the vehicle and walking, removing rubble from a doorway and going through the door, climbing a ladder, using a tool to cut a hole in a wall, opening valves, and pulling a fire hose a short distance and connecting it to a standpipe.
During the finals next year, Pratt said DARPA will combine all the tasks into a more authentic mission for disaster response.
“We’re trying to move it from an engineer’s test to an authentic test,” he added, “and we’re trying to do that at just the right rate so that it matches the maturity of the technology.”
Still, Pratt said, “I think we can make the finals a little harder than I had thought.”
No matter how the finals turn out, the robot technology will have to be commercialized before it can be used in disaster response and mitigation, he said.
“What the commercial sector can do is to find a market for the technology. It may not be in disaster response, it may be in health care, agriculture or manufacturing. We just don’t know,” Pratt added.
“But the most important thing is, after DARPA does its work, we count on the commercial world to reduce the cost, find the markets and produce things of value,” Pratt explained, and only afterward can the disaster-response community put the robots to work in those applications.
According to its DRC fact sheet, DARPA fully expects all those things to happen.
“Technologies resulting from the DRC will transform the field of robotics,” the fact sheet says, “and catapult forward development of robots featuring task-level autonomy that can operate in the hazardous, degraded conditions common in disaster zones.”
And of the very successful Florida DRC trials, DARPA’s Tousley said, “Every time DARPA does a challenge we learn something new and we try and make that process better, and even better for the development of technology for the future.”