Other changes from the trials include that the eight tasks must be performed in sequence, so that getting stuck on any one task disqualifies the team from performing the remaining tasks. Also less information will be given out about the tasks, and one task will be a complete surprise, which no team will be notified about ahead of time. Four identical courses will be set up at Fairplex so that simultaneous runs can be made. Each team will be given two chances, with their best score counting, and the surprise task will be different on each day of the two-day event. Ties will be broken by the total time elapsed performing their tasks.
DARPA started with 16 teams at the December 2013 trials, which included tasks such as driving a vehicle and cutting hole in wall. Eleven of those 16 teams were prequalified to participate in the finals, however one -- Team Schaft (Japan) was bought by Google and has since dropped out but Team Thor from Viginia Tech made up the difference by splitting into two teams, Team Valor which remains at Virginia Tech and the new Team THOR at the University of California. At least 13 self-funded teams, including three new teams sponsored by the European Union and the governments of Japan and Korea, are expected to be added by submitting video tapes (in February 2015) proving their robots can complete the basic tasks needed, for approximately 24 teams competing for the $2 million purse in June 2015.
The 11 finalists so far are:
IHMC Robotics (Florida Institute for Human & Machine Cognition, Pensacola, Florida)
Tartan Rescue (Carnegie Mellon University, National Robotics Engineering Center, Pittsburgh, Pennsylvania)
Team MIT (Massachusetts Institute of Technology, Cambridge, Mass.)
Team WPI-CMU (formerly Team WRECS, Worcester Polytechnic Institute, Worcester, Mass.)
Team Trooper (Lockheed Martin Advanced Technology Laboratories, Cherry Hill, NJ)
Team ViGIR (TORC Robotics and Virginia Tech, Blacksburg, Va.; TU Darmstadt, Germany; Oregon State University, Corvallis)
Team THOR (University of California, Los Angeles)
Team Valor (Virginia Tech, Blacksburg, Va.)
Team KAIST (Daejeon Metro City, Republic of Korea)
The finals will be open to the public on June 15 through June 16, 2015 in the Fairplex, Pomona, Calif. with stands set up for at least 2,000 spectators. After the finals, DARPA expects to take the winning teams on a world tour performing exhibition events.
The total budget DARPA is spending on the Robotics Challenge is $95 million.
Also don't forget that the Robot Olympics will run alongside the 2020 Olympic Summer Games (for humans), which will be hosted in Tokyo.
— R. Colin Johnson, Advanced Technology Editor, EE Times
Yes, raising the bar will definitely lead to new innovations. In fact, I would not say:It would be fascinating, but that: It will be fascinating to see what technology developments come out of this competition.
This competition is grueling and it is great to see DARPA continuing to raise the bar. It would be fascinating to see what technology developments come out of thsi competition that actually get commercialized.
Yes, you are right about DARPA ofter failing--usually because they set their expectations to high. So it is refreshing that they set them high, but not high enough. There is worldwide interest in deveoping smart robots--one reason--and of course a $2 million purse doesn't hurt :)
@DrQuine, there is an implicit assumption in your prediction - that it will be a robot instead of many. That is really where the complexity lies. Break out the tasks and imagine a robot for each and the problem becomes much simpler. This is similar to the advent of electric motors. The predictions at the time were that the complexity of bringing all the work to a motor would be overwhelming. The solution: Build a whole bunch of motors.
@KB3001 Good point. Personal care of the elderly is a growing need. It will be interesting to see how the humanoid robot gets "qualified" (certified) as a registered nurse, LPN, or aide. Once done, they will have the advantage that a given model of robot may be precertified without the need for individual attendence in years of nursing school as humans are required to do. I'd predict a very long development cycle before they're able to do much more than deliver meals. They'll certainly need solid programming to ensure fail-safe operations and appropriate responses to unexpected problems.
Not sure about your "robots" prediction, DrQuine. The Japanese are already introducing humanoids to help the aged. Most advanced countries will follow suit because they will face the same problems soon.
I'm pleased that DARPA is making such good progress that the bar can be raised. All too often, projects fail to meet their objectiive and the expectations must be lowered. Congratulations to DARPA and the hard working engineering teams.
I predict that we will not see many robots in daily living in the future. They will be concentrated in the military, heavy manufracturing, and emergency rescue applications. Instead I expect that the Internet will become even more ubiquitious (as if it isn't already wiith SmartPhones). smart devices (like autonomous cars) will require less human intervention, and we'll all be struggling when failed devices (intermittent problems can't get fixed) make it hard to do things that used to be simple.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.