PORTLAND, Ore.—The future of search-and-rescue robots will be charted over the next two years by a new "grand challenge" set to the most advanced robotics researchers in industry and academia. Eighteen teams from around the world will compete to design the most autonomous search and rescue robot for assisting humans during natural disasters in the Defense Advanced Research Projects Agency (DARPA) Robotics Challenge.
Next week the newly chosen teams will prove out their paper designs in a virtual arena that mimics the real-world arena in which the robots will compete later in 2013, with the final contest—including a $2 million purse—slated for 2014.
DARPA cites the Fukishima Dai-ichi nuclear meltdown in Japan as its motivation for bumping Humanitarian Assistance and Disaster Relief (HADR) into a top priority for robotics at the Department of Defense (DoD). The Naval Research Laboratory already has prototype on-ship firefighting robots in humanoid form—so they can navigate passages and use firefighting tools designed for humans—and the DoD already has remote-controlled robots for sifting through rubble for survivors. However, the grand challenge set for the DARPA Robotics Challenge (DRC) is fully autonomous operation in environments originally designed for humans, but which have been damaged beyond repair by disasters. Key will be the ability to navigate human rubble-strewn passageways, use any human tools they find there as well as being able to follow instructions from humans who had had no prior training in robotics.
DARPA chose the Willow Garage spin-off the Open Source Robotics Foundation (OSRF) to build its official arena simulator for the first phase of the competition, the Virtual Robotics Challenge (VRC) being held next week (June 17-to-27, 2013).
The arena for the Defense Advance Research Projects Agency (DARPA) Virtual Robotic Challenge (VRC) was designed for it by Willow Garage spinoff the the Open Source Robotics Foundation (OSRF).
"As a non-profit neutral steward of the robotics community, we've created an open source ecosystem that is attractive to robotics companies the way open-source Linux/Appache/Pearl-Python is attractive to Internet companies," said Brian Gerkey, OSRF founder.
The OSRF simulator will allow contestants to test out their designs in a virtual environment where they can evaluate whether their prototype will meets DARPA's contest goals before building them in hardware. Robots that pass muster virtually will then be built-to-spec by their developers for the real-world trials to be held in December. In all, 18 teams will participate in the VRC next week, but only the 11 winners will progress to the prototype phase for the initial hardware competition in Decenber 2013. The final contest will advance the top eight teams to the December 2014 finals where the winner will be awarded a $2 million purse.
The top seven teams, each of which received $1.8 million for development costs, consists of Carnegie Mellon University, Drexel University, NASA's Jet Propulsion Lab, NASA's Johnson Space Center, Raytheon Schaft Inc. (Japan) and Virginia Tech. And 11 teams receiving $375,000 in development funding including Ben-Gurion University of the Negev, Carnegie Mellon, Florida Institute for Human and Machine Cognition (IHMC), Lockheed Martin - Advanced Technology Laboratories, Massachusetts Institute of Technology (MIT), NASA Jet Propulsion Laboratory (JPL team two), Robotics Engineering Excellence, Team ViGIR, TRACLabs Inc., University of Kansas and University of Washington.
From 18 VRC contestants, five winning teams will be chosen from the seven receiving $1.8 million today, to receive an additional $1.2 million in funding, each, to develop their hardware prototype for the December 2013 contest. Six teams from the 11 receiving $375,000 today will be chosen from the VRC winners to receive an additional $750,000 to build their hardware prototype for December. From those 11 teams progressing to the December 2013 hardware prototype contest, eight winners will be chosen to receiver an additional $1 million each to further develop their prototype for the final competition in December 2014 and its $2 million purse.
My most recent robotic invention is a "robotizing" implant inserted into the brain of a human. Sort of like the "Borg", only more robotic and much less tribal. The biggest advantage is in the range of motions and in the coordination. The downside is that the strength is not as great.
It may very well be that no single robot can live up to the challenge. But, one nice thing about robots is that we can make allowances for that. I wonder if it would be within the rules to have a "carrier" of sorts. One big robot that has a few separable smaller specialized robots. It could have the little snake bot for going through small holes or maybe a quadcopter for flying over impassible debris.
DARPA is certainly triggering many creative projects with their ongoing challenges. I wonder how "out of the box" solutions fare under such guidelines. Will the "search and rescue snake" that crawls through small holes have a chance against a human sized robot that can walk through doorways and carry out a victim? They each solve important problems.
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.