DARPA launches $2,000,000 Robotics Challenge
DARPA has launched its Robotics Challenge, in which teams will compete to develop a robot that can assist humans at disaster sites
Attention, BEAR, MABEL, and SAFFiR: You may be able to win your makers two million bucks! That’s the amount that DARPA is offering to the victorious team in its recently-announced Robotics Challenge. The winning robot will be the one that best meets a series of challenges, designed to test its ability to provide assistance in disaster scenarios.
The contest is open to anyone, be it robotics manufacturers, universities, or even basement tinkerers. In fact, DARPA (an agency of the U.S. Department of Defense) is encouraging the involvement of people not normally involved with robotics in order to bring fresh perspectives to the field. It is hoped that the event will foster innovations in both hardware and software.
“This challenge is going to test supervised autonomy in perception and decision-making, mounted and dismounted mobility, dexterity, strength and endurance in an environment designed for human use but degraded due to a disaster,” said DARPA program manager Gill Pratt. “Adaptability is also essential, because we don’t know where the next disaster will strike. The key to successfully completing this challenge requires adaptable robots with the ability to use available human tools, from hand tools to vehicles.”
Scheduled to start this October, the first phase of the contest will stretch over 15 months. This will be followed by a second phase, running throughout 2014. Both phases will culminate in a competition, at least one of which will likely require a single robot (per team) to carry out the following eight tasks:
- Drive a utility vehicle at the site – this will include getting in and out of an unmodified vehicle, along with steering, accelerating and braking on a moderately curving road (maximum speed will be 15 km/h, or 9.3 mph)
- Travel dismounted across rubble – after leaving the vehicle, robots will have to traverse a variety of terrains, along with avoiding insurmountable obstacles
- Remove debris blocking an entryway – that debris will likely be a solid object such as a rock or cinder block, not weighing more than 5 kg (11 lbs)
- Open a door and enter a building
- Climb an industrial ladder and traverse an industrial walkway
- Use a power tool to break through a concrete panel – the tool will probably be an air or electric impact hammer and chisel, or an electric reciprocating saw
- Locate and close a valve near a leaking pipe – there will be multiple pipes present, but the leaking one should be identifiable by its hissing sound and escaping smoke
- Replace a component such as a cooling pump – besides locating the pump, this will involve loosening one or more fasteners, extracting the faulty pump, then reversing the steps to replace it with another
The PETMAN humanoid robot (Photo: Boston Dynamics)
Teams will have the option of competing using the GFE (Government Furnished Equipment) Platform, designed and built by DARPA. This will be based on Boston Dynamic’s bipedal ATLAS robot – which is itself derived from the company’s PETMAN – and will be a basic humanoid robotic platform including arms, legs, a torso, and head. Teams using a GFE will be free to modify its hardware and software as needed, and will be aided in that process with a computer test-bed known as the GFE Simulator.
That said, organizers have stated that the robots do not have to be humanoid in form ... although some of the tasks certainly suggest that humanoid robots might be particularly well-suited to the challenge.
Source: DARPA via IEEE Spectrum
About the Author
An experienced freelance writer, videographer and television producer, Ben's interest in all forms of innovation is particularly fanatical when it comes to human-powered transportation, film-making gear, environmentally-friendly technologies and anything that's designed to go underwater. He lives in Edmonton, Alberta, where he spends a lot of time going over the handlebars of his mountain bike, hanging out in off-leash parks, and wishing the Pacific Ocean wasn't so far away.
All articles by Ben Coxworth
If I were them, I'd build a remote controlled mech - basically giving a visual/optical (and possible haptic) feedback to an actual person positioned in a suit somewhere, who - from there - could then control the humanoid.
Problems with this approach would be responstime, and (as always) balance. The latter could be solved by making the making the humanoid crawl on all four; of course, it/he should be able to stand up, and even walk (e.g. to get into the car) - the crawling should just be to avoid falling too much, thereby saving time.
(that the humanoid crawls doesn't mean that the person should also lie down; that's the beauty of not actually being there.)
Hmmm really brilliant - but hey it's DARPA!
Just think, after you build it and win it, you will end up having to make a million more of them, that are better, faster, heavily armed and armored - to defend against the fully weaponised ones they send against you.
Its darpa people. They gonna put guns on it! Didnt anyone watch terminator? Its skynet all over again!
We will (and now do need) robots like these to plant a trillion fruit trees! Especially in hostile climates and conditions. Remember - for the evolution of sustainable life on Earth we need an alternative means of productivity that can do ecological farming and set up designer ecosystems. there is always risk of famime , war , poverty when humans control the Earth alone. New robots could not inevitably support problems forever -that is not logical - so what is the problem except that the focus of all you readers should be to support and petition for robots that can save us from environmental hazards and do the work load that is needed for ecological farming. Contact me if you can help me get robots to plant fruit trees soon.
MasterG....when did Skynet happen?
'....along with avoiding insurmountable obstacles...' I guess insurmountable obstacles won't be so insurmountable after all? Or does all objects need to be mounted? Pffft....
This is a very dumb idea. Or should I say extremely premature? The tech to actually build something like this is way off. The first and biggest problem is powering such a device/robot/machine. Not going to happen. Also, why does it need to be able to drive? Self-driving car technology already exists and will hit mainstream in the next 10 years. I'm not saying it's going to take over or be big initially, but it will be at available at the consumer level.
This request basically is to build an almost fully operational human in robot form. Good luck....but don't be surprised when you fail miserable. Most of these pie in the sky projects fail miserably. It's good thought to push the envelope of technology, but I think this is a little ridiculous.
1) survey the area
2) establish the whereabouts of inhabitants
3) find Sarah Conner
Having participated in the first "Robotics Grand Challenge" put on by DARPA, I have to applaud their continuing efforts. To all the naysayers: Yes, the first Grand Challenge was a "dismal failure" but was followed by unqualified success in the next two Challenges! The "prize" system is how great leaps in technology occur most efficiently. Yes, DARPA is part of the DOD but have you ever heard of "spin-offs"? The "can't be done" types have always been beaten by the "let's try it" crowd. Will we get a practical 'humanoid' robot out of this first try? Probably not but they will learn so much from "failure" that the next try will be a resounding success. The only "problem" is how the use of humanoid robots will affect the human workforce and economics. This has already been envisioned by the SciFi authors and dealt with fairly well.
"Failure" is only failure if you quit. Engineering, medicine and all other human endeavor is only advance by "failure" because that fosters inovation.
"Only the Unknowing can believe that Mankind's future is restricted to one small, overstrained planent in a Galaxy of plenty; The Meek shall inherit the Earth, the rest will inherit a Universe"-Unknown
History nut what an awesome quote im stealing it from you! Skynet cos even the bible warns us of this, in the koran it warns us of this too - dajaal (anti christ) has one eye that sees and a smaller glowing red eye (seen a cctv camera lately?) Yes lets give robots a gun and lets see how long before the amazing darpa Ai (another darpa project) takes before it sees us virus like human beans as the THREAT. Robots dont need fruit trees, warmongers need murder death kills, darpa dont do fruit trees.
I'm just dying to see the some competing robots =)
And to all of you naysayers: Have you ever thought about, where the internet came from?
DARPA - is the one of the best things that has ever happened to America.
This is a good idea. Just imagine the spin-off from this in time. If you were around at the start of the twenty-century and told people then that in time almost everyone one day will have these horseless carriage (the car) you would get the same remarks everyone today are make about domestic robots to come. I see only good will come from this. I recon in time we will all have one and trade it in for the next new model as we do with the car, each one better than the last.
Remember the automobile is not cheap, yet we think it is worth spend such large amount of hard earn salary on it, because of how is saves time and gives us freedom to travel at will when we want to. I recon in time robots will be the next most important thing we will have in our homes for the same reasons above and more, such as, when we are disable or just old, it will carry out it tasks twenty four seven. That in its self will give another new life to us all when we enter that part of our life.
Yes, when I think of these things, it would be a wonderful thing to invest in one day. It would be to me what my car is to me now and would give me the freedom when old that I have now when middle age. Roll on with the R&D. I welcome it.
If you are in the military (or a close relative), when do you start asking "why didn't they send a robot in, instead of me/them"? Who makes the decision whether to sacrifice a robot or a human? What are their priorities? Cost???
Be positive about it. Imagining the winning robot with guns and weapons may be a bit too drastic. The purpose is to create 'robot that can assist humans at disaster sites'. Just think of walle :)
The problem with your concept is that the activities of the test require autonomy. In other words, the robots have to be able to do it themselves. What you are describing is a machine being operated by remote control, and we already have that in abundance.
Such advances as this test envisions would not only allow independently operating robots to operate in extremely dangerous conditions in which they could be saving lives, but also for establishing human living conditions on other planets. Imagine a team of robots assembling living quarters on the moon or mars in advance of the arrival of human astronauts.
Hello, first artificial intelligent is not advanced enough to make a terminator. Second this contest looks more like an attempt to make a previous robotic body function instead of coming up with a new humanoid robot.
I have been able to make the robot you want for a number of years.
I designed and prototyped a completely dexterous prosthetic hand 20 years ago. It is strong, lightweight and could pick up a dime or handel thread. recently I found an old device that simplifies the drive construction. I also designed and prototyped a subminiature flexible linear encoder; great for a digital control glove and very fine position control. I am presently searching for the company that makes a hydraulic actuator using a piezoelectric pump, very high strength.
telepresence makes most all the programming unnecessary.
As far as putting weapons on them, why? My design could do anything a human could do, only better and have haptic feedback. A few other parts make it truly human like movements possible.
Over 160,000 people receive our email newsletter
See the stories that matter in your inbox every morning