Official Details For the DARPA Robotics Challenge 61
An anonymous reader writes "The DARPA Robotics Challenge is offering tens of million of dollars in funding to teams from anywhere in the world to build robots capable of performing complex mobility and manipulation tasks such as walking over rubble and operating power tools. It all will culminate in an audacious competition with robots driving trucks, breaking through walls, and attempting to perform repairs in a simulated industrial-disaster setting. The winner takes all: a $2 million cash prize."
DARPA is awesome (Score:1, Informative)
No other government agency inspires me like DARPA does.
"Show me a piece of future technology" (Score:1)
That's it? (Score:2)
Re: (Score:2)
Re:That's it? (Score:4, Insightful)
Re: (Score:3)
Hell, we don't need the government for adventurous, useless robotic machines.
Purdue University runs a contest every year [wired.com].
Re: (Score:2)
You give unlimited rights to DARPA (and the US govt), by participating in this contest. I am pretty sure the US govt would be willing to pay a lot more if you were to sell robots that perform these tasks.
Re: (Score:2)
There is a danger they would hand the plans off to a different contractor to have them build it, but I think you could make the case that your company woul
Re: (Score:1)
Per the PDF at the bottom of https://www.fbo.gov/index?s=opportunity&mode=form&id=ee8e770bcfe1fe217472342c67d6bd5a&tab=core&_cview=0
"DARPA desires Unlimited Rights, as defined in DFARS 252.227-7013 and
-7014,12 to all deliverables generated by the DARPA Robotics Challenge performer under this
effort except clearly-identified, widely-available, commercial software tools, with their
commercial availability described and substantiated in the proposal."
"2.4 Intellectual Property {1}
Per section VIII
Re: (Score:2)
These are the same rules that generally apply with all government research. Unlimited rights does not mean that they own the copyrights, patents, etc.. It just means that they can take your data to use it internally for there own purposes. One reason for this is that there are a lot of small companies that can't build up the manufacturing capability to meet the demands of the government. Just FYI, there are a few very very rich defense companies that have managed to do just fine with these same rules.
Re: (Score:2)
That 2 million is more of a performance bonus rather than a contract.
Wait a minute... (Score:1)
robots driving trucks, breaking through walls, and attempting to perform repairs in a simulated industrial-disaster setting.
I think I saw all that in 1984 [imdb.com].
Immenant Disaster... (Score:4, Interesting)
This is classic DARPA... while these grand challenges are good for focusing research initiatives, they tend to ask for much more than the field can offer in a reasonable amount of time given the funds. Look at the first grand challenge: not a single team finished the race, and even the best team from the best school with the most funding only finished 12km of the proposed 240km course.
It was an utter embarrassment. Only after they relaxed the requirements in 2005 of the competition to more accurately reflect what was humanly (or more aptly roboticly) possible at the time was the competition a success.
Now they're expecting a full-on humanoid that can drive a car, bust down walls, move rubble, operate tools, all in unstructured environments? Look at the DARPA ARM grand challenge, where the state of the art could barely do these kinds of manipulations in a controlled well-lit laboratory.
On the other hand, I suppose if they're allowing teleoperation/assisted autonomy that makes things a lot easier. I guess I just don't want a repeat of the collective embarrassment of the robotics community that happened in 2004.
Re: (Score:3)
RTFM - the robot does not have to be humanoid, it simply needs to be compatible with human tools.
Re: (Score:3)
Did you RTFA? The competition will involve much more than that: it needs to drive a vehicle, including entering it, using all the controls as-is, and exiting the vehicle; it needs to climb ladders and traverse grated catwalks; it needs to use tools made specifically for humans; it needs to traverse several types of terrain, including moving over rubble and then move said rubble.... exactly what kind of robot would you propose that can do all that?
If it is not a humanoid it is going to be highly engineered
Re: (Score:2)
I did read part of the FM :) My comment wasn't intended to suggest that any robot of non-humanoid shape would meet the requirements, just that humanoid shape isn't a requirement - only the illustration.
You are correct - the requirements are stiff. Many of the challenges lend themselves better to a multi-legged chassis with manipulators that can handle tools. The human form is tricky for a number of reasons, odd center of gravity, etc. that are more suited to biological form than a machine. This looks lik
Re: (Score:3)
http://www.youtube.com/results?search_query=dlr+justin [youtube.com]
The ask is not unreasonable at all. Just requires focus, and investment. Prizes generally generate much more research dollars put into the thing than the prize value, so it's a super good investment.
Re: (Score:2)
And compatible with a drivers seat in a vehicle that has not been modified to accept the robot.
It will need a fairly human shape to fit in the drivers seat, it will need fairly human upper limbs to use the controls and fairly human legs to operate the pedals.
Re: (Score:2)
Damnit you're right.
Here is video of an advanced prototype from the 80's modifying a vehicle for use.
http://www.youtube.com/watch?v=4TBcQ8h_kXU [youtube.com]
Can you just imagine what they have now?
Re: (Score:2)
Re:Immenant Disaster... (Score:4, Interesting)
And in round 2, it worked. (Score:2)
I guess I just don't want a repeat of the collective embarrassment of the robotics community that happened in 2004.
Yes, the 2004 Grand Challenge was a disaster. I was there to watch. In the 2005 Grand Challenge, all 23 teams that made it to the California Motor Speedway had better systems than anyone had in 2004. I had a team there. Yes, the course in 2005 was easier, but some of the vehicles there could have completed the 2004 course.
I don't think the humanoid challenge will succeed on the first try, either, since the schedule is so tight. The first trial with real robots is only 15 months away. By try 2 or 3, tho
Re: (Score:2)
Relaxed the requirements how? As I recall the second year course was more difficult than the first one.
Looks like a job for... (Score:2)
1. Execute complex tasks in dangerous, degraded, human-engineered environments." Check! (I assume Detroit counts.)
2. Supervised autonomy. Check!
3. Mounted mobility, dismounted mobility . Check!
4. Dexterity, strength, and platform endurance. Check!
Does an ACL reconstruction count as "robotic"?
Does the robot that breaks through walls... (Score:3)
...have to yell "OH YEAH"! and be red...
Re: (Score:1)
DARPA: Precipitating tomorrow's dystopia today. (Score:1, Funny)
Re: (Score:1)
That's about a week in duration at DARPA. You can be the proudest fellow around until the next installment of Dancing with the Stars is broadcast. That's certainly worth a big pat on the back.
Re: (Score:2)
Slashdot, where yesterday's technophiles become today's luddites.
Re: (Score:2)
Re: (Score:1)
Re: (Score:1)
Re: (Score:2)
Build one that can break through bunker walls ... (Score:2)
. . . and the Defense Department will pay you more than $2 million.
. . . Israel is rumored to be acquiring large amounts of clay and letters on parchment for a similar project appropriately code named as "Golem."
. . . and North Korea's missile launch will be foiled, when a Mechagodzilla springs out of the ocean and steals the missile, after emitting consuming a giant jar of Kimchi and flatulating the audience.
and when we have autonomous vehicles... (Score:1)