NASA Needs Prize Contest Ideas

Michael Huang writes "If you like the idea of tech contests--think ANSARI X PRIZE and DARPA Grand Challenge--and you also like space, then NASA wants you. It needs ideas (and rules) for the Centennial Challenges, prize contests with $20 million funding in 2005. Current ideas (download Excel spreadsheet) include: Mars and asteroid microspacecraft missions, lunar robotic landing, robotic triathalon, rover survivor, Antarctic rover traverse and extreme environment computer. Wikipedia has good coverage."

AI is not ready indeed: with online demo.

[ControlFreal]

Parent: +6 Insightful

I'm a researcher in AI, and I can do nothing but backup the parent's claim (sad as it is). When we use AI, we would expect a robot to be able to perceive its surroundings (analyzing sensory inputs), make decisions (reasoning) and act (generating actuator outputs).

I can only comment on the first, since I'm a Ph.D. student in Computer Vision. And the general picture is, to be quite honest, depressing. Forget all you've seen in e.g. Terminator (e.g. the robot analyzing its visual input, and all the nice text in the image): it ain't gonna happen for a long time! Although space missions are (presumably) less complex in terms of sensory inputs, the state of affairs in dealing with normal natural images gives a nice idea of what's currently (im)possible:

I'll provide an example here. I'm doing Computer Vision (face-detections), and the current state of affairs is about this: When finding faces in 800x600 images, this can be done in about 1 second (yes: 1 full second), at about a 90% detection rate and a couple of false detections per image. For more complex object classes that are not so nicely symmetric (think cars, houses, landscapes, etc.), the performance is dramatically worse.

You can look at the BitTorrent link [unimaas.nl]. And ONLY if that doesn't work, use this [unimaas.nl]. As for reasoning: this is still in it's infancy, but I'm not working in that field, so I cannot comment on that well. Any takers? ;)

Re:Practical Long Lasting Space Suit

[VernonNemitz]

Yes skinsuits have been studied and prototyped. Jerry Pournelle wrote about them years ago in his "A Step Farther Out" column.

Re:THIS FALL ON NASATV!

[BKDotCom]

NBC had that in the works and payed $40 million for the rights for "Destination MIR"

http://www.cnn.com/2000/TECH/space/09/12/mir.sur vi vor.ap/

Workshop on June 15-16 in DC, open to public

[FleaPlus]

As noted on the web page, there's the Centennial Challenges program is organizing a workshop [nasa.gov] in Washington DC on June 15-16. You can register online [tisconferences.com] (please, authentic registrations only!). This sounds like an excellent opportunity to help shape history.

Here's the blurb from the web site [nasa.gov]:

To kickoff Centennial Challenges, NASA's new program of prize contests, NASA will host a workshop on June 15-16 in Washington, DC. The purpose of the workshop is to:

1. Gather ideas for Challenges,
2. Develop rules for specific Challenges and gauge competitor interest in various potential Challenges, and
3. Promote competitor teaming.

This workshop will be a key input into Centennial Challenges planning, helping to determine what specific Challenge competitions NASA announces in 2004 and 2005 and the rules of those competitions. All potential Centennial Challenge competitors, including interested members of industry, academia, students, and the general public, are invited to attend.

Re:Practical Long Lasting Space Suit

[DerekLyons]

Unfortunately, I can't find any references to skinsuits, although I gather they've been tried in prototype --- can anyone confirm this?

You can't find any references because the largely don't exist other than as a sci-fi staple. They have numerous significant problems, and no clear paths to solving them. (Go to Google Groups and search for 'skinsuit' or 'skinsuits' in the sci.space.* groups. They are a perennial topic of discussion there.)

Here is a pretty good reference!

[VernonNemitz]

.PDF format [usra.edu]