Darpa Robotics Challenge Results: Terminator's Grandpa Crowned
After 33 months of herculean effort, DARPA’s Robotic Challenge has finally ended with South Korea’s HUBO robot coming out on top followed by the United States’ Running Man and Chimp.
This whole 3 year robo-olympic competition had it’s genesis ultimately linked to the tragic Fukushima earthquake in April 2011. That earthquake and resulting tsunami heavily damaged the Fukushima nuclear reactor causing 3 of its 6 reactors to meltdown. This event joined Chernobyl as the only time we have ever reached a level 7 on the International Nuclear Event Scale.
At some point soon after the disaster, workers entered one of the reactors to manually vent hydrogen. They walked into an increasingly deadly wall of radiation causing them to beat a very hasty and worrisome retreat. Later, the unvented hydrogen contributed to massive explosions which further damaged the reactor and contributed more radioactive materials to an already contaminated environment.
The Japanese emergency teams realized that radiation resistant robots could have prevented the worsening of the crisis but they had mothballed and then gave away their working prototypes years earlier since they were deemed unnecessary. How ironic that a robo-topia like Japan was left with its pants down for want of robots. Not only Japan however, there were no robots on the planet that could do much more than go in and take selfies with the damaged reactors. This was a wake-up call for the robotics industry and now many governments realize the need for autonomous robots to aid in disasters like Fukushima.
This is where DARPA comes in. They began the DRC or Darpa Robotics Challenge to spur innovation and get roboticists collaborating. Apparently, the robotics industry was ripe for such an intervention. Dr. Gill Pratt, Program Manager DARPA Robotics Challenge said:
“DARPA is in the innovation business, not in the development business. So, what we do is we wait for technology to be almost ready for something big to happen, and then we add a focused effort to catalyze the something. It doesn’t mean that we take it all the way into a system that’s deployed or to the marketplace. We rely on the commercial sector to do that. But we provide the impetus, the extra push the technology needs to do that.”
The entire DRC challenge itself was in three stages over three years beginning with an entirely virtual component. The best teams were then filtered into the next hardware stage. Finally after dozens of teams competed in these earlier stages, the best 24 international teams made it to the last stage this past June 5th and 6th.
Instead of separate 30-minute tasks, this finale consisted of a single hour to sequentially accomplish 8 tasks. These tasks simulated, appropriately enough, a disaster at an industrial plant:
- drive a small car
- get out
- open a door and pass through
- close a valve
- cut a hole in the wall with a drill
- either climb over debris or clear a path
- climb a flight of stairs.
- A surprise task
When this Octatholon was done, only three robots were able to complete all 8 tasks. The winners of the gold/silver/bronze medals ( actually: 2 million, 1 million, 1/2 million US dollars) were:
- Hubo of Team Kaist of Daejeon, Republic of Korea (44 minutes)
- Running Man of Team IHMC Robotics of Pensacola, Fla. (50 minutes)
- Chimp of team Tartan Rescue of Pittsburgh, PA (55 minutes)
Couple interesting things about the robots. Many looked similar because the robotic platforms were donated meaning they didn’t have to be created from scratch. Seven teams used the Boston Dynamic Atlas robot and 2 used the Korean-made Hubo platform. This made it less of a hardware competition than a software and control systems competition. Scott Lavalley, one of the design leads at Boston Dynamics said of the Atlas robots:
“They’re all the same…They can modify it with sensors and computing. It’s all controls, all software [which differentiates the teams]. We deliver the hardware, and then it’s up to them to write software and navigate it through the course.”
Not all were donated of course and there were definitely some obvious differences between some of the robots. Hubo may have had one of best innovations in the form of wheels on its knees. By simply kneeling it could roll on fairly flat surfaces making it very stable. When actual walking is required, it simply stands up to take the stairs, do the Cha Cha, whatever.
If you’re wondering about the whole humanoid robot vs non-humanoid robot philosophies, there is no clear winner at this point. The bottom line is that humanoids were fast but not very stable for some of the tasks. Conversely, the non-humanoids were more stable but clearly had problems in areas that were meant for humans.
And finally, is it just me or did anyone else imagine Data from Star Trek TNG going thru this course and finishing it with a flourish in about 5.2 minutes?
Video: DRC robots falling down
Image Credit: https://wtvox.com/2015/06/team-kaist-wins-the-darpa-robotics-challenge-2015/