7 Things We Learned After the DARPA Robotics Challenge Finals
The DARPA Robotics Challenge Finals have come and gone. Here are some of the most important takeaways from the ultimate robot competition.
Now that the DARPA Robotics Challenge Finals are over, it’s time to take a look at some of the things we learned while Team KAIST enjoys its $2 million grand prize.
1. Our Expectations Were Too High
There’s been a lot of talk about how underwhelming the DRC Finals were. Myself included. How come each robot couldn’t complete all eight tasks flawlessly? Look how often the robots fell. Why did it take the winning team so long (44 minutes, 28 seconds) to complete the obstacle course?
But at the end of the day, that’s probably not a fair conclusion to have reached. Yes, there was a lot of money, time, and effort spent on developing these robots, but these were some of the best minds in robotics at work. I was perusing some of the comments on Reddit and came across reader GreyMX who makes some great points:
“These robots have successfully completed a challenge that literally no robot in the world could have done three years ago. It has also spurned public interest in the broader potential applications of robotics, and we’re likely to finally start seeing robots being useful outside of factories as a result.
“If you were underwhelmed, then I can only imagine that you set literally impossible standards for what you were expecting from the competition. The top teams at the DRC finals consisted mostly of the world’s leading experts in bleeding edge robotics, so it’s not as if the money could have been offered to better candidates.”
The capabilities of robots will continue to improve, and competitions like these are part of the reason why.
2. DRC-Hubo Should be Baseline Robot for Future Challenges
Team KAIST from South Korea dominated the competition, performing all eight tasks flawlessly in a time of 44 minutes, 28 seconds to win the $2 million grand prize. The second-place finisher, Team IHMC Robotics, completed eight tasks in 50 minutes, 26 seconds, and Carnegie Mellon University¹s Tartan Rescue Team came in third with a time of 55 minutes, 15 seconds.
Team KAIST’s DRC-Hubo cuts a hole in a wall during the DARPA Robotics Challenge Finals.
So why was the DRC-Hubo so dominant? Erico Guizzo and Evan Ackerman of IEEE Spectrum have a fantastic piece that details everything that went into Team KAIST’s performance. The team “custom designed and built almost every part of the robot,” including motorized wheels on both of DRC-Hubo’s knees, a rotating torso to spin its waist up to 180 degrees, extra long 7 degrees-of-freedom arms, a simplified vision system and much more.
DRC-Hubo was by far the most advanced robot there - it won by nearly six minutes. If the competition is truly meant to find robots to aid in disaster situations, let’s give the best robot available to the industry’s brightest minds and let them go to work.
3. Where Are All the Female Roboticists?
This is perhaps the most alarming (or maybe not) thing you’ll learn about the DRC Finals: about 95% of the teams are comprised of males. In total, only 23 of the 444 members are women. Eleven out of the 24 teams were all male. Eleven of the 24 teams are made up entirely of men. The University of Tokyo, for example, has 43 members and zero women.
Team KAIST celebrates winning the DARPA Robotics Challenge Finals and $2 million.
What’s causing the lack of female roboticists? Are robots too male looking and acting? Our sister publication, Robotics Business Review, takes an in-depth look at this problem.
4. Walking is Still Tough for Robots
Just three out of 23 teams managed to complete all eight tasks set for the robots, which included driving and exiting a vehicle, opening and going through a door, locating and opening a valve, using a tool to cut a hole in a wall, removing an electrical plug from a socket and putting it in a different socket, traversing rubble and climbing stairs.
There’s a fantastic compilation (watch the video above) of the best robot falls from Day 1 of the two-day competition, and most of them occurred with the robots trying to stand or walk. But a lot of robots fell multiple times, but thankfully there wasn’t significant damage. Teams could call for a reset if their robot fell, but it would come with a 10-minute penalty.
5. Robot Competitions Are Here to Stay
No matter what your opinion is of the DRC Finals, robot competitions aren’t going anywhere anytime soon. There are some serious issues - aging populations, natural disasters, dwindling natural resources - that robotics can certainly help.
There are robot competitions happening all the time all over the world. Some big. Some small. Japan is hosting the Robot Olympics in 2020. The European Union has 100 robot projects currently running with over 700 partners, committing 70 million to 80 million Euro a year to fund new projects.
And as DARPA Director Dr. Arati Prabhakar put it, “we aren’t in competition with each other. Really we are in competition against the problems of the world.”
6. Robot Autonomy Remains Low
“The state of autonomy in robotics these days is actually only beginning, and so the level of autonomy we’re talking about is at the task level, not at the mission level,” says DARPA program manager Dr. Gill Pratt.
Robots that can open a door on their own without a human operator need to have humans that tell it what to do beforehand. The robot can then open the door and then wait for the next communication from humans.
7. Robot Uprising Still Far Off
Here are some of the phrases uttered to describe this DARPA robotics challenge:
- “It’s like watching paint dry”
- “It’s like watching grass grow”
- “It’s like watching a golf match”
I enjoy golf, so I’m not sure what that’s all about, but you get the point. Certainly progress has been made, and the performance by the DRC-Hubo was impressive, but the thought of a robot uprising seems like science fiction after the DRC Finals.