Forgot your password?
typodupeerror
Robotics Hardware

DARPA Grand Challenge A Real Race At Last? 173

Posted by Zonk
from the flat-out dept.
museumpeace writes "News.com has posted its second story in a week on a hopeful Grand Challenge contender. Stanford's Stanley, a VW Touareg run by 100,000 lines of code can hit 40 mph and has now traversed all but 3 miles of last year's desert course without problems. A few days earlier, Carnegie Mellon University's Team Red announced that its Sandstorm, a modified Hummer, had run 200 miles without any problems though on a closed track. DARPA cut the field to 40 in June and will cut it to 20 before the race in October."
This discussion has been archived. No new comments can be posted.

DARPA Grand Challenge A Real Race At Last?

Comments Filter:
  • by Realistic_Dragon (655151) on Tuesday July 19, 2005 @04:54PM (#13107886) Homepage
    I wish my girlfriend could drive 200 miles without crashing into something. Perhaps theres a trade in programme or something?
  • It's spelled Touareg. Could VW have come up with a more confusing name for its premium 'ute? I still don't know how to say it!
  • by RandoX (828285) on Tuesday July 19, 2005 @04:56PM (#13107903)
    Was it refueled on the fly?
  • by Anonymous Coward on Tuesday July 19, 2005 @04:57PM (#13107911)
    How many miles per line of code?
  • Princeton (Score:4, Informative)

    by Adrilla (830520) * on Tuesday July 19, 2005 @04:59PM (#13107925) Homepage
    A guy from princeton made some posts claiming that his team had run the entire course and then some last night. They can be found in this thread [slashdot.org] ps: they haven't made the final roster yet.
    • I know two of the six members of the Princeton team, and their entry sounds quite promising despite lacking the infinite budget of my own school's entry (CMU). And while conquering basic controls problems including obstacle avoidance, I think the real problem as usual will be obstacle detection, not obstacle avoidance. As I was last informed the team uses only stereo cameras and image processing. It will be interesting to see how they can handle the low contrast environment of the desert as well as havin
      • Oh and as far as "the entire course" they mean a test course not the GC course which isn't released until 2 hours before the challenge itself.

        If I remember correctly, they were referring to last year's course.
  • is that to be impressive?? if it was like 500 i would be WOW even say 5000 but 100000. is that jsut the main driver or is that each part sumed.. and when do we get to read the doc it should be intresting // this line was added after it saw the tele pole as a lane line
    • Actually... the way the article phrased it:

      The computer scientists have written more than 100,000 lines of code to tell Stanley what to do.

      Made it sound like the journalistas thought that was a *lot* of code. I was just thinking that didn't sound like much at all, given the team-size and project scope. Maybe 10,000,000 or 100,000,000 would be a lot. 100,000? That's about a typical mid-sized software project. (Depending on the environment of course)
    • Impressive...if each line consists of 1,000,000 characters...
  • by pudding7 (584715) on Tuesday July 19, 2005 @05:00PM (#13107948)
    I was out at California Speedway last year when they were practicing/qualifying or whatever.

    Boring as hell. Interesting, but very tedious.

    Machine moves 10 feet, stops.
    Moves 5 feet, stops.
    Moves 40 feet, stops.
    Turns 10 degrees.
    Turns 15 degrees.
    Moves 10 feet, stops.

    Hopefully they've picked up the pace a bit. Otherwise they'll never go 200 miles through the open desert in the alloted time.
    • Whittaker described the start of the race as ENFORCED boredom for safety sake..they have a speed limit off the line because thats where the crowd will be...funny huh? a race with a speed limit.
      • funny huh? a race with a speed limit.

        Maybe it's being sponsored by C. Montgomery Burns? [snpp.com] :o)
      • Which is quite unfortunate because those V10 TDI Touareg's are really fricking fast and powerful. The horsepower is similar to the V8 (gas) but it has DOUBLE to torque (and both of them at lower RPM). They are quite the the little monster, I work as a runner at the largets VW service center in the country and drove my first one a few days ago (the v10 is fairly rare) and boy...I dont like SUV's at all but...that thing was amazing. It also was VERY cleanburning and VERY fuel efficient

        Unfortunately it lo

    • Hopefully they've picked up the pace a bit. Otherwise they'll never go 200 miles through the open desert in the alloted time.

      What does it mean when someone says they ran "last year's course" without a problem?
      Does the car take to the road knowing nothing of the obstacles ahead or has it been programmed to avoid the hazards revealed in a year's study of the route?

      • in order to make it to the qualifying round, your vehicle must show path following abilities and obstacle avoidance. the darpa guys tested this during the site visits by placing some obstacles on the way points (gps coordinates) and some near way points. so, to answer your question: it should handle arbitrary, unknown obstacles. that said, they may well incorporate a priori knowledge of the course into their control.
    • Machine moves 10 feet, stops.
      Moves 5 feet, stops.
      Moves 40 feet, stops.
      Turns 10 degrees.
      Turns 15 degrees.
      Moves 10 feet, stops.
      Apparently it took 100,000 lines of code because it's written in Logo [wikipedia.org].
      FORWARD 5
      FORWARD 40
      RIGHT 10
      RIGHT 15
      FORWARD 10
  • by American AC in Paris (230456) * on Tuesday July 19, 2005 @05:02PM (#13107973) Homepage
    Standford's Stanley, a VW Taureg run by 100000 lines of code can hit 40 mph

    I'll bet the teams from Harverd, Berkely and NIT are quaking in their boots.

    • by Anonymous Coward
      I'll bet the teams from Harverd, Berkely and NIT are quaking in their boots.

      Do you mean MIT?
      • by Anonymous Coward
        Captain, I'm detecting extremely low levels of clue in the area.
    • a VW Taureg run by 100000 lines of code can hit 40 mph

      How much lines of code can it transmit when loaded with tapes and speeding down the highway?
      • Re:Yes but... (Score:2, Insightful)

        by f0dder (570496)
        Hope they have code to fix breakdowns.. Poor volkswagon.. a check of consumer reports and the '04 Toureg gets a predicted reliability of Solid Black (poor for those who don't read CS). But good owner satisfaction (white circle). A Ford Exploder comes out as good reliability (White circle) and fair owner satisfaction (dark half circle). A Toyota 4 Runner on the other hand has Very good Reliabiity (red half circle) and excellent Owner Satisfaction (filled red circle)
    • Go Cardinals!
  • This is, after all, funded by DARPA so I wonder what creepy guys are drooling over the results in the shadows?

    Myabe some will be rejected from the competition because there is no possible way to mount .50 caliber machine guns on the vehicle.

    Now of course there are practical applications to this too, like an ambulance that can really haul ass and save lives, or a taxi that will take the most direct route. A pizza wagon that can bake and deliver tasty eats. I am sure other /.'ers can think of some positiv
    • The idea is not to develop a military machine. The idea is to develop the technology that could then be applied to a military machine. There is no expectation to mount a .50 acl to the top of any of these machines.

      Can you imagine a Volkswagen with a big ass machine gun on top of it?
    • a taxi that will take the most direct route You've obviously never driven a taxi. Taking the most direct route results in the lowest fares! Experienced taxi drivers learn to quickly spot out-of-towners, then take the most roundabout route possible, thereby maximize their income!

      Nope, not good for ambulances, which usually operate in high-population density areas. Even experienced ambulance drivers have accidents; I wouldn't trust a machine to do this. Nope, wouldn't work for taxis; this would require sophi

    • by CrankyFool (680025) on Tuesday July 19, 2005 @05:59PM (#13108489)
      I don't think you're paranoid -- I think you're ignorant.

      Last I looked, the ratio of support personnel to fighting personnel was something like 7:1 in the modern military. That means that for every person whose primary job is actually killing people, there are seven people whose primary job involves support functions. One of the big support functions is "get this stuff from point A to point B."

      While sometimes "get supplies from point A to point B" can be rather hazardous -- our supply units in Iraq can talk about that -- it can also quite often be boring, repetitive, and relatively safe. Being able to send supplies to the front lines without having to equip a truck with people will help alleviate this need.

      The DoD spends many billions of dollars on research every year. The really sexy research -- "how do we kill people better, faster, and in more bulk?" -- gets the most coverage, but a very significant chunk of their research is around things that are not directly related to the whole "kill or be kill" thing -- for example, the internet.
      • Wow, not even one word about the human targets of such a system?

        Obviously not much of a concern in defensive operations, say an invasion of America by an enemy army. But frankly we don't do much of that, do we?

      • Exactly. Well said.

        Why, it seems like just yesterday [slashdot.org] I was telling someone something similar.

        I have actually talked to one of the handful of DARPA people who hatched this idea in the first place.

        Interestingly, the primary motivation for this is for cargo and supply-line applications.

        I am not saying that it couldn't be used for ground-based unmanned attack vehicles eventually, clearly it could. Eventually. But that kind of use would require a much smarter and more flexible maneuvering capabil

    • How long will it take for Foster-Miller [foster-miller.com] to implement the software from whatever team wins this thing into their Talon [globalsecurity.org] system...?
    • This is, after all, funded by DARPA so I wonder what creepy guys are drooling over the results in the shadows? Myabe some will be rejected from the competition because there is no possible way to mount .50 caliber machine guns on the vehicle

      Of course it will have military uses, it is sponsored by Darpa and the Congress has already decided that one third of all military vehicles must be autonomous by 20XX (sorry, forgot).

      And why would it be so terrible if they mounted a .50 cal on it? Regardless of yo
      • Its the simple fact that its expendable.
        Nowadays, at least the coffins comming back from iraq reminds people a little bit that waging isnt a game.

        How would it be if you just paradrop smart vehicles (or remote) and just play "enemy combatant extermination". Nobody gets hurt (at least no "real people", i.e. americans) no unsighly people crying about fallen family members... That way war would really be fun, right out of every neo-cons wet dream.

        Thats whats worse.
        • How would it be if you just paradrop smart vehicles (or remote) and just play "enemy combatant extermination".

          The thing I don't get about this line of argument against unmanned combat vehicles is that we already have killing devices which we can paradrop over other people with approximately zero risk to our own soldiers. They're called bombs.
      • Unless the mission is "go here, and shoot everything that moves", we're not going to see that for a long time.
  • by nizo (81281) * on Tuesday July 19, 2005 @05:04PM (#13107982) Homepage Journal
    When do they start allowing contestant cars to carry guns/rockets/etc to take out other cars?

    In fact, this could be a whole new brand of reality show: a few dozen death row inmates are released, followed minutes later by "smart cars with guns" that chase them down. Last inmate alive gets a pardon.

    • Wish I had mod points today, but I used them all yesterday!

      They should take the top 3 vehicles that finish the course, add weapons and turn them loose again. The one that comes back wins.
    • That'd be sooo awesome. It could be like that old Apple 2e AutoAssault game. Or maybe it'd be more like Road Rash.
    • The show's working title is Twisted Metal.* Bob Barker will host the show, complete with reused The Price is Right theme song. For the matches, Darrell Waltrip will provide play-by-play, John Madden will give color commentary, and there will be occasional features narrated by Defense Secretary Donald Rumsfeld. Insiders say Scott Peterson may star in the first episode.

      *No word on whether clown-driven ice-cream trucks will be allowed.

  • No gas? (Score:5, Funny)

    by slapout (93640) on Tuesday July 19, 2005 @05:06PM (#13108003)
    run by 100000 lines of code

    Wow. A car that runs on computer code instead of gas? That's great! Now I can program myself home.

    (Wonder how I'll pay attention to the road while I write code though....)
    • Sure it runs on code, thats easy to figure out.

      It is bloatware, so it eventually rots. Then the gas is fermented in tanks and passed to the engine and is burned.

      Imagine what kind of mileage a windoze car would get.
    • I resubmitted this to get it posted at all,,,the /. eds LOVE concision and they snipped this from the post:
      Robotics legend Red Whittaker heads Team Red, his former student, Sebastian Thrun, is QB for the Standford team....
      Gentlemen, boot your navigation computers!
    • run by 100000 lines of code
      Wow. A car that runs on computer code instead of gas?

      Uhh, no. Try again.

      It clearly states that it's run *BY* code, not *ON* code.

      Unless you're suggesting that in your car, the gasoline performs the navigation?
    • run by 100000 lines of code

      Can I have that number in standard Libraries of Congress units, please?
  • Try navigating rush hour traffic...now that would be a trick. Could have a robotic hand for giving someone the bird....just a thought!
  • by martin_b1sh0p (673005) on Tuesday July 19, 2005 @05:12PM (#13108068)
    VW must love the publicity. First they set a world record [auto123.com] and now this. If they beat a Hummer I'll bet the sales & marketing dept will have a field day !
    • They also had a pretty damn strong showing in the last couple years of the Dakar rally.

      Personally, I think the touareg has way too many "control modules" but in the case of the car being computer controlled, maybe its a good thing that everything is already computer accessable but its just...the first time I sat in one (I am currently a car runner at a VW dealership), the seat and steering wheel started moving as soon as I got in the car--damn luxury power seat package thing--and of course the person we

  • by tlambert (566799) on Tuesday July 19, 2005 @05:13PM (#13108083)
    20? What are the elimination criteria for the vehicles DARPA's simply not going to let compete?

    Specifically, if no one has ever done this successfully before, how the heck do they know what a successful approach looks like?

    I understand dropping the obvious non-starters - teams whose vehicles crash or get lost on a small test course, or teams whose vehicles are not ready to go at all - both of these are valid rejection criteria.

    But it seems really silly to set an abitrary number at "exactly 20"; the article doesn't really explain how the decision on whether or not your vehicle "makes the cut", other than "was evaluated by DARPA experts" - who have yet to solve the problem themselves.

    -- Terry
  • by eSims (723865) on Tuesday July 19, 2005 @05:25PM (#13108206) Homepage
    Standford's Stanley, a VW Taureg run by 100000 lines of code can hit 40 mph and has now traversed all but 3 miles of last year's desert course without problems.

    Herbie!

  • the "news" articles i pasted together to make this post were, in part, PR for two of the better funded competitors...they WANT to pique interest and get publicity for the event. A few disclosures that a sponsored vehicle has demonstrated a capacity to do what none of the entrants did last year is sort of infomercial. If the race were more like real racing [would NASCAR be an example?] the teams would keep mum about what they were capable of to throw off the odds makers. I think the smartest comments I ha
  • Too Bad (Score:2, Informative)

    by techsoldaten (309296) *
    I think it's too bad we can construct a rover that can charge around the landscape of Mars for months at a time but cannot get a Taureg or a Hummer to drive through the desert on it's own.

    Different disciplines, I know, but still. It would just seem like someone should be able to put something together that could overcome any obstacle and proceed in a straight line from point to point in the race. Like a Bradley tank, maybe with some modifications that exchange armor for nitrous tanks or something.

    I realiz
    • by rucs_hack (784150) on Tuesday July 19, 2005 @06:06PM (#13108553)
      The principle challenge here is effective navigation of difficult terrain. The Mars rovers are extremely slow, manually driven, (albeit by extreme time delay), and it takes a *lot* of skilled people to keep one going. If autonomous vehicle technology is significantly improved, then the effect on the the mars exploration program would be vast. And actually the desert scenario is pretty close to the problems found on mars: Sandstorms rocks sand more sand holes Imagine Mars being explored with ten (or more) completelly autonomous and robust vehicles scooting around looking for interesting features. That would rock.
    • The Caltech Team is reasonably close to various people at JPL and as such they were all set to give us their code, but DARPA told us we couldn't use it (no big surprise there). It's not really a big deal since it's a better learning experience for our team to develop the code themselves.

      On a separate note, I believe that our vehicle is the only one that can fit a reasonable number of people inside, comfortably, with seatbelts (6 I think).

      • That's because you choose a huge ass SUV for your vehicle. Obviously people with smaller vehicles will have less seating. The MITRE Meteor for instance only seats two because it's a Ford Ranger.
        • I suppose I should clarify. Yes there are teams that chose small vehicles but there are plenty of other teams who chose large SUVs (a certain H2 comes to mind) and still have little seating space. I'd be damned impressed if you managed to have 4-5 seats in your Ford Ranger after installing a bunch of computers and onboard sensors.

          Anyway, its more of an aside, because by the time the military gets around to implementing this technology, someone will come up with an elegant way of arranging all of the co

  • >a VW Taureg

    it's called "Touareg"

    it's not even a german word so there is no reason to spell it that wrong

    http://www.vw.com/touareg/index.html [vw.com]
    • Well, there may be no reason to spell it wrong, but I'd be damned if I've heard anyone pronounce it correctly. It's got the opposite problem most Russian words are endowed with (too many vowels).

      What language does "Touareg" come from anyway and what the smegma does it mean?
    • Why do I need cookies turned on to look at pictures of their cars?

      Fuck VW.
  • UK Grand Challenge (Score:2, Interesting)

    by Warfire (412710)
    I was part of a team at my UK university that built a robot inspired by the Darpa Grand Challenge. We ran off battery power and theorised we could go for 5-6 hours on twin 16Ah lead acid batteries. We moved at a steady fast walk so unscientifically we moved at 3mph... so maybe we could have covered 18 miles on the batteries we carried. We didn't look into fuel power because of the low budget we had, £1,500 rather than the 10's of thousands in the Darpa teams budgets.

    Here's a picture from our local
  • Just wondering... (Score:2, Insightful)

    by Mad Ogre (564694)
    If these contestants realize that the winning robot will pretty much go straight into military R&D. Not that there is anything wrong with that. It just smacks of "Real Genius". If a robot could do this out on the battle field, we could save a lot of lives just in ambulance roles alone. Other applications abound, but just think of it... just in an automated ambulance roll... the robot rushes in, get's loaded up with casualties, races back to the field hospital, all without concern for enemy fire. Tha
    • If a robot car came out of nowhere and yanked me out of my nice warm foxhole and then zigzagged through an active battlefield with bullets flying everywhere, you bet your ass I would be concerned.
  • All Mac solution (Score:3, Informative)

    by no_opinion (148098) on Tuesday July 19, 2005 @06:30PM (#13108755)
    A friend of mine is captain of Team Banzai [gobanzai.com], and they're building the car's brains out of Macs and using iSight cameras for vision. The car is a VW Touareg.

  • More info (Score:3, Informative)

    by heli0 (659560) on Tuesday July 19, 2005 @06:46PM (#13108910)
    http://www.tomshardware.com/game/20050713/index.ht ml [tomshardware.com]

    Good article about the Stanford and Cyberrider teams, with video.

    The project is based on a partnership of several companies. Volkswagen provides the vehicle, a diesel-based Touareg R5 TDI with an automatic transmission, and the automotive know-how. Intel is supplying the Pentium-M chips (Dothan core) as well as the platform consisting of several 855GM-based mini-ATX boards. Then there is MDV (Mohr Davidow), a company that invests in start-up firms. Android, Honeywell, Tyzx and Coverity are among the other sponsors. Tyzx has specialized in image recognition software, which is especially important for AI (artificial intelligence) systems.

    It goes without saying that the power hungry P4 processors and the corresponding platform would have made a bad choice. After all, the power dissipation loss of these seven systems would easily have exceeded 1200 Watts! This would have required special modifications to be made to the alternator and several other components. Instead, Intel has supplied the project with its tried and tested Pentium M notebook processors. These use the Dothan core with its 2MB of cache, are clocked at 1.6GHz and run on an 855GM platform. As a result, the overall power requirements of all seven systems together are only 280 Watts. The individual nodes are linked using a Gigabit Ethernet network.
  • If you want to see some cool videos and read more about the actual design of the CMU entries (Sandstorm and H1ghlander) check out their website. The race log [redteamracing.org] section is especially interesting.
  • All I can say is, WTF?

    The software runs on six Pentium M processors, which are Intel-made, low-power chips originally designed for the telecommunications industry.


    Is this story written by an idiot? Or did he need to stick a few more words in there to make it longer? Or is this some sort of Intel sponsored text-based product-placement?
  • And the software now uses machine-learning technology to program the car to "remember" how it's first driven over a course manually, and then emulate those actions for autonomous driving.

    First a human drives the course and then the software replays the actions of the human? Yes, the software makes corrections along the way but this defeats the purpose of the competition, i.e. the car autonomously navigating by itself over unknown terrain. Sounds dangerously close to a disqualification to me (if I were

"The whole problem with the world is that fools and fanatics are always so certain of themselves, but wiser people so full of doubts." -- Bertrand Russell

Working...