NASA Prizes for Builder and Flyer Robots

FleaPlus writes "NASA has recently announced a couple more X-Prize-style Centennial Challenges. The first is a Telerobotic Construction Challenge, for using a team of robots to assemble structures from building blocks with minimal human intervention. The second is an Planetary Unmanned Aerial Vehicle Challenge, to create a robot which can fly a path using visual navigation and hit ground targets with a probe (no GPS allowed). Rules are still being finalized, with the contests scheduled for 2007. Both prizes are for $250,000, the max Congress is allowing NASA to offer."

Comments

250K?!?!?!

[Punboy]

These tasks are probably much more complex than the Ansari X prize... which rewarded 40 times this much. Offering $250K is insane. Stupid. Insanely Stupid.

Re:250K?!?!?!

[ichigo 2.0]

According to Babelfish:
It was once sweet rabbit. The rabbit had a large carrot. It was a special carrot: if you ate a piece of it, then she regenerated within fewer minutes again. The rabbit hid it in its earth hole. But a beautiful daily, when the rabbit came back to its hole, was it away. The rabbit frightened and began a hole to dig. It dug and dug to it on a large cave pushed. The carrot is because of the other end of the cave. But there was also an orange, petrified cow with wings.


Orange petrified cow with wings, indeed. :)

Fixed prize limit?

[Bombula]

$250k is a not a whole lot of money. I'm not sure how many outfits would be able to get something out of the design stages without more money than that, so this prize would most likely not even cover costs.

That may not be the point, but it would sure be nice to at least have the development costs for projects like this covered by prize money.

Re:Fixed prize limit?

[massivefoot]

I think the unmanned aerial vehicle might be doable for that cost. I mean, the hard bit is the AI. Other than that you've just got a model aeroplane being controlled internally rather than be someone on the ground with a transmitter.

Of course you need to know how far it has to fly etc, but IMO it can't cost over £10k to build the airframe. The programme to fly it wouldn't be too hard either, the only hard part is that it needs to know where it is and what orientation it's in . Yhe latter is trivial - use a giro like any other aircraft would. The former could probably be done by taking either a stereo image from two cameras mounted on the wing tips, or useing some sort of downward looking radar, to create a 3D picture of the local terrain, which could then be matched to a map held in memory and a "best-fit" obtained.

Re:Fixed prize limit?

[Somatic]

Four words: Chief Knock-a-Homer.

Re:Fixed prize limit?

[massivefoot]

Well their prize only requires that your plane fly on Earth. However, you've raised some interesting points, I suggest you go to www.x-plane.com and download the demo. It lets you fly on Mars (or at least the previous versions did, I assume the new one also does). There are two aircraft included in it that can fly on Mars, but they handle very oddly and you're right, you need huge wings.

Re:Fixed prize limit?

[acaspis]

it can't cost over £10k to build the airframe

Extra points if your design scales to the low-density atmosphere of Mars, and can fly slow enough to do the probe thing.

the only hard part is that it needs to know where it is and what orientation it's in

Well, that pretty much sums up one of the most painful problems in robotics.

Yhe latter is trivial - use a giro like any other aircraft would

No. Gyros drift. Aircraft autopilots rely on other things, like radio beacons on the ground, or GPS. Or the sun.

The former could probably be done by taking either a stereo image from two cameras mounted on the wing tips

To get accurate 3D information out of this, you'd need to be flying pretty close to the ground.

or useing some sort of downward looking radar,

Yes, lookup "IFSARE". Good luck tracking your elevation data while flying over a flat plain or lake - then you need the visual clues too.

AC

Re:Fixed prize limit?

[xeoron]

The point perhaps: To have a reasonable starting point, then there it is all about tweeking and evolving....

Re:Fixed prize limit?

[Skyfire]

I do a fair amount of work with small UAVS (10-20 foot wingspan), and a 2-stroke powered propeller is definetely the way to go. Its been done for decades with success, its far more fuel efficient than small gas turbines (For subsonic speeds, all propulsion systems benefit from economies of scale. In general, the bigger the area of air being accelerated, the more efficient it is. Thats why the turbofans on modern airliners (B-777) are so large.), and parts are cheap and readily available. I question the ability to build a quality UAV to carry that much weight for that distance though. We are talking about 500$ for engine, minimum 5 servos x 80$ each, and that only gives you 100$ or so for the airframe. I would estimate for the basic airplane around 2000$ or less.

Matching funds?

[Bombula]

Since the $250k limit is imposed by congress, maybe matching funds could be sought from private sponsors. Surely some of the big contractors like Boeing and those sorts of guys would be willing to put up some prize money if they might end up getting the big contracts to develop and build the real things?

Should be interesting

[inflex]

While the prize money isn't a lot, it might actually have an interesting effect in bringing about the less well funded but no less ingenious developers into the foray while the typical high ranking developers/companies pay less attention. It's certainly something I'd be interested in participating in (the aerial navigation).

Hmm...

[St0rmwarden]

I wonder if GTX Global will be entering anything with their "true AI" onboard...

Already something like the second one:

[OverlordQ]

The Association for Unmanned Vehicle Systems International holds an annual competition named the International Aerial Robotics Competition. In order to win, a team's vehicle must complete these four objectives autonomously, without human intervention:

        * Fly three kilometers along a specified path of GPS waypoints.
        * Search a target area for a marked building; identify the openings and their centers.
        * Enter the building and document specific aspects within its interior.
        * Complete the previous three goals consecutively in less than 15 minutes.

Yes it's GPS vs Visual, but roughly similar

Re:Already something like the second one:

[Alterscape]

I was involved in my university's AUVSI/IAR team two years ago. I didn't have enough CS background to participate in the development, but I did a lot of work building up and modifying the ARF airframe we purchased for our competition use (that mainly consisted of building new higher-lift airfoils and making mounting arrangements for our camera and control system). I attended meetings and talked to the programmers and from what I know, GPS was almost invaluable in our solution. You can achieve kind-of-the-same sort of accuracy with an inertial guidance system and a known take-off point (which is what US fighter aircraft did for non-radio navigation in the pre-GPS days) but that lacks the same precision, and isn't nearly as simple as plugging in a GPS receiver that spits out coordinates in an immediately-useful fashion.

So the no-GPS thing -is- a real difference between the competitions.

An aside: When I was involved, we didn't have to actually fly -into- the building. This sounds like it tilts the playing field very far in favor of helicopters or other VTOL solutions.. but that's not relevant to the NASA prize.

Military?

[Mr2cents]

I have just read the robotic competition faq yesterday, and I remember a similar competition there.. I just found it back at http://avdil.gtri.gatech.edu/AUVS/IARCLaunchPoint. html. I was just looking for a fun competition so that I have some fixed requirements for building a robot myself, but it's either too advanced or too simple. If anyone knows of a fun competition in Europe, please let me know.

"Fully autonomous ingress of 3km to an urban area, locate a particular structure from among many, identify all of the true openings in the correct structure, fly in or send in a sensor that can find one of three targets and relay video or still photographs back 3km to the origin in under 15 minutes."

It looks similar, although the prize money is only $50k, and it's for military use.

Knock, knock. Reality?

[Anonymous Coward]

Is the goverment totally attempting to demonstrate how disconnected from reality it has become? I'm trying to understand the reasoning behind these ridiculous terms: Do something incredible hard, mostly for free, no attempt on our part to even pretend to cover a fraction of the cost with the reward money. I'm sorry, why? The goverment hands out billions in Corporate welfare, funds global police actions, but this is their idea of a reasonable way of conducting themselves in some sort of scientifc endevour?

I am a high school student

[PlayfullyClever]

I am a high school student, so I believe I am qualified to answer you.

First, be forwarned. I don't mean to sound cynical, but there is not a whole lot that has to do with science and technology that would excite most students. Even if it does, a lot of people are too scared of being called a "nerd" or a "geek" and thereby having their social status for the rest of the four years ruined to show that excitement.

There are, however, some. I don't think that a robotics competition is a good idea, however. I don't know about most schools, but at mine there are not a lot of people interested in robotics. Besides, it would take a lot of work, and a lot of the most brilliant people are inherently lazy.

I think the programming fair was a great idea, however. Every time I write a program to do the simplest thing on my TI-83+ graphing calculator (such as convert celsius to fahrenheit for instance) people gape at me with awe and amazement and ask, how did you DO that? This includes jocks, socialites, and various other groups of people who would normally not be caught dead showing an interest in the "nerdy" fields of computers or technology.

If you put on a programming fair, you are not going to be able to teach anyone computer programming in a day, but you will spark their interest. Give away a few CDs with C tutorials on them or something, and maybe, just maybe, a few kids will try them out.

Also, expect the bit-head population to turn out in force at your fair. You can even put some of them to good use, having them help the newbies who have no idea what's going on.

In conclusion, programming fair=good, robotics competition=bad.

Re:I am a high school student

[Ada_Rules]

If you put on a programming fair, you are not going to be able to teach anyone computer programming in a day, but you will spark their interest. Give away a few CDs with C tutorials on them or something, and maybe, just maybe, a few kids will try them out.

Give high school students CDs with C tutorials on them...Ugg..Why don't we just give them crack pipes. I've seen the damage a few months of C programming can do to a yound mind..It is not pretty.

Milltary connection?

[FhnuZoag]

The second is an Planetary Unmanned Aerial Vehicle Challenge, to create a robot which can fly a path using visual navigation and hit ground targets with a probe (no GPS allowed).

There is obviously a military connection here. For probe, read bomb, or bullet. Essentially, it's designing the next generation of autonomous UAVs. Presumeably, our military planners now believe GPS to be possible to compromise in times of war. (fairly reasonable thanks to the new attention on space war)

What are the ethics of this sort of competition?

would you use fundable.org?

[swframe]

The results could be open to the public.

I have a shover robot, if they need it

[jovlinger]

It's great for pushing things