Sunday, January 08, 2006

Meditations of a Gadget Freak

I've been fascinated by the DARPA Grand Challenge, won in October by a team from Stanford. The January issue of Scientific American has an article, Innovations from a Robot Rally, which is a must for every red-blooded gadget freak, it goes into detail about how the various teams solved their technical issues. (The link above is to the article on the SciAm site, but you could just go buy the magazine. In fact, if you want to see the photos and diagrams, you have to get the magazine; it's only $11 - in the U.S. - and they're worth it.) Between March 2004 and October 2005 these teams went from no vehicle even completing the course (I don't think any entry got farther than 3 miles), to 4 vehicles doing 132 miles in under 8 hours: the winning time was 6 hours 53 minutes. That's an average speed of just over 19 MPH.

Some of the most interesting tweaks weren't even done by the winning team:

One of the obvious problems was, how do you track where the robot is? The obvious answer is, give it a GPS, or maybe several; but how do you track distance and direction if it gets into a canyon where the GPS signal is blocked? The team of high school students from Palos Verdes, CA (yes, they competed) developed something they call a GroundMouse, based on the principle behind optical mice: with a camera, a bright light, and an optical tube, they built a 2 dimensional odometer with 1 millimeter accuracy in any horizontal direction. (The team staffed by professional engineers built something similar, but they used a Doppler radar. Isn't money nice?) I love 2 things about these kids: first, that they tried at all. Second, that their solution used stuff you could buy at Radio Shack. There's hope for the next generation yet!

The Indy Robot Racing Team solved a problem that won't even (really) exist until autonomous vehicles go commercial, if they ever do, but without this solution or something like it, they never will: these guys built a network protocol for swapping "plug and play" sensors and software modules in and out of their vehicle. They built the interface. Them's my kind of engineers! (Disclosure: I used to be a systems programmer.)

These are not the only fascinating technical solutions: you have to see the picture of the one with the bank of 64 lasers on the vehicle roof, on a motorized circular platform, spinning 10 times a second, and backed by a bank of signal processors programmed in Assembler. Obstacle detection: could spot something the size of a person at 500 feet.

I strongly recommend any fellow techno-weenies out there to take a look at this article.

8 comments:

  1. Anonymous5:54 PM

    hedera, I read about this in various news reports. The positive offshoot from this would be the automobile that navagates by itself. But you know DARPA will not use the resulting technology in an enlightned way. This is John Poindexter's old agency and they'll just use the technology to kill people. Sorry to be a downer.

    your pal, cooper

    ReplyDelete
  2. cooper, you're right, of course, that what DARPA wants is something they can use on a battlefield without risking troops. U.S. troops, that is. On the other hand, remember that DARPA funded the development that led to the Internet: what eventually comes out of a DARPA effort isn't always what DARPA thought it would get.

    I haven't seen anything that indicates that any of the technology developed for the Grand Challenge has been classified. Certainly none of the stuff written up in Scientific American was classified. This means it's out there, and all the people who worked on it know about it; and that means if one of the teams thinks they can make some money off it, off they go into the capitalist jungle. Autonomously navigating, of course.

    I was particularly thinking about the high school team, and the long term impact on those kids of competing with the pros in an effort like this. As long as the urge (and the ability!) to create gadgets isn't dead, this country has a future.

    If you want to get really deep, you can ask: just because this might be used in an unpleasant way (i.e. to kill people), does that mean we should not work on the project? At that point you're getting into, we should not invent things that can be used for evil, even if they can also be used for good. This is much more ambiguous than the Manhattan Project, since I can't think of a single non-compromised application for a fission bomb...

    ReplyDelete
  3. Anonymous6:16 PM

    hedera, what are you, some kind of brainiac or something?

    Okay, we both have valid points and I don't see where we're really disagreeing that much. I can see the value of the high school kids participating in solving an engineering problem. I'm certainly not a Luddite. I work with kids high school age and younger through the Rocketry club I belong to, teaching them the basics of Physics as it applies to the hobby. Technology has always been a double edge sword, with good and not so good applications. But don't DARPA sponsored this competition to be win the Good Citizenship Award of 2005. It hoped to get a variety of good solutions to the problem and, also, hoped to recruit a couple of high tech heavy hitters in the bargain, pulling them over to the dark side.

    ReplyDelete
  4. Anonymous6:21 PM

    Oops. "But don't 'think' DARPA..." proofread your copy, boy!

    ReplyDelete
  5. Anonymous5:48 PM

    hedera, so how's the leg these days? Have you gone dancing, yet? Back at work? Oh let me guess, they loaned you a wireless notebook, so you could work from your near-death bed. Those bastards!

    ReplyDelete
  6. No, cooper, I don't have a wireless laptop; my employer is extremely paranoid about wireless security and discourages the use of it. I'm still off, still doing rehab - I finally got my disability claim approved yesterday (don't ask) and will return to work on 2/22.

    I'm not dancing but then I never did dance much. I walked a mile today; I need to do that more often. I'm now doing my water aerobics twice a week again - very carefully and about half speed.

    My biggest problem is that I want to be better, soon, and I push too hard, and then I have to stop and quit aching. My husband says I only have two speeds, "on" and "off"... I think 3 days of rehab with weights and 2 water aerobics may be too much for a week, I need to pace myself better.

    ReplyDelete
  7. Anonymous8:19 AM

    hedera, I somehow suffered a pinched nerve at the base of my spine (L5-S1, for the medical types out there) on Easter vacation. My left leg went numb and I drug it around like a trailer with a flat tire for about 6 months. Many tests and specialists later, the neurosurgeon said, "I don't know what is wrong with you, but your nerve is completely dead, you'll just have to live with it and I can't help at all."

    So, Plan B was launched - At my wife's insistence, I went to a Chiropractor for the first time in my life. One look at the X-ray and he said, "I think I can help." Well, long story short, "IT'S A MIRACLE!" Working with him and a neuromuscular massage therapist, I'm back to about 80% function and getting better every week.

    Now, I said all that so I could tell you don't push the rehab too much. At least with my injury, that only makes it worse.

    Best of luck. I'm not much of a dancer either, much to my wife's disappointment.

    ReplyDelete
  8. cooper, no matter how bad things get, I guess there's always someone in worse shape! What a horrific experience! My personal opinion of chiropractors is equivocal, but I never argue with success; if it's working for you, go with it.

    My goal is to be able to do a 5 mile hike; I guess I need to quit trying to do that by next week... It appals me how fast you lose muscle. Before my surgery I was bouncing up and down the stairs in my house multiple times a day (arthritis and all; I used a cane). Now, it's all I can do to climb them (17 steps), I've lost that much leg muscle. Yikes! But it's dogged as does it, and I'll keep on.

    ReplyDelete