Slashdot Mirror
Carnegie Mellon Wins Urban Challenge
ThinkingInBinary writes "The results from the Urban Challenge are in! Carnegie Mellon's Tartan Racing team came in first (earning a $2 million prize), followed by Stanford's Stanford Racing team in second (earning $1 mil) and Virginia Tech's Victor Tango in third (earning $500k). Cornell's Team Cornell, University of Pennsylvania and Lehigh University's Ben Franklin Racing Team, and MIT, also finished the race in that order."
2nd is the best
So these guys get some millions from public funding and does the public get any opensource out of it ?
And will it blend?
Ginga no Rekshiya Mata Each page.
MIT, MIT...
Oh yeah, isn't that kind of like Massachusetts' version of CMU?
While the immediate winners of the race are the three teams holding checks, as well as the military which gets to pick from a field of highly successful new technology, the real beneficiaries will be the drivers of the world. I believe the importance of this hasn't quite filtered into most people's minds.
Many people know that more than 40,000 people die each year in motor vehicle accidents, however when it comes to people I feel this number is insufficient. "More than 40,000 people" have been dying each year now for more than a decade, and that's only in the US. Since I was 17 more than four hundred thousand people have died participating in an activity that machines can now do flawlessly (if very slowly). This blows my mind.
Worldwide, 1.2 million people die on the roads every year and the repercussions of these deaths on families and friends can be unusually devastating due to their sudden, unexpected nature.
The performance of these three teams is akin to three major pharmaceuticals all announcing they have come up with a cure for one of the major cancers. That, surely, would have been worldwide front-page news.
Now, of course, the real debate begins. How much more will consumers be willing to pay for safe vehicles, and what limitations on speed will they accept? Rolling out this technology (if you'll excuse the play on words) will require changes in infrastructure, law, and cultural mentality. Especially here in the states. If it means saving this many lives, will you pay twice as much and drive at half speed, at least for a little while?
How far this technology has come in just a few years is (ridiculously) amazing. Major kudos to everyone who's brought this so far!
I only wish that one of the conditions of winning was to release the software that powered your car - can you imagine how much farther things would have come if everyone could build on the previous years' winners? So much brilliant coding has gone into this, but so much of it is just reinventing the wheel. (...Ouch.) But in all honesty, the state of the art would progress gigantically if one of the winners would GPL their car-driving software.
Actually, according to the Wired blog, MIT came in fourth, although the other teams were not mentioned.
Nothing at all in that summary tells me what the Urban Challenge is; nothing in ANY of the links tells me concisely what it is, either; Wiki eventually did. How hard would it be to include "a prize competition for driverless cars" in the first sentence of that article?
Are y'all experimenting with automated posting or something, because that at least would make sense.
Triv
It's kind of interesting how much effort has gone in to building a robot that can drive in (error-prone) human traffic. If, on the other hand, *every* car was automated, it would be so much easier to implement. (Controls built into the road, maybe, and of course less need to handle wildly out-of-control cars; plus benefits like optimized freeways (anyone remember "Blue Thunder"'s freeway?) and intelligent intersections that talk to incoming cars, etc.) I think the eventual progression is to automated and efficient public transportation, where no one owns their own car, nor needs to. Did anybody consider, back in the day, if one car per person/family was actually a good idea?
Driving is a privilage, not a right.
I don't think its about that per se. It's about finding and implementing the most efficient transportation mechanism. If you could develop a fully automated system, you wouldn't need to own cars since they could be available on demand. How many hours are cars driven vs garaged, one could reduce the total number of automobiles by a factor of 5 if not more.
I remember seeing an article on here a while ago about mass transit that went to each neighborhood but instead of trains were 4 passenger vehicles that were fully automated.
Ahhhh.. *lightbulb*
Carnegie Mellon's algorithm
//crossing an intersection
if(OtherCars.SignallingToCross())
{
Me.Stop();
Me.WaitForClear();
}
OshKosh Truck's modified algorithm (copied)
//crossing an intersection
if(OtherCars.SignallingToCross())
{
//Me.Stop();
//Me.WaitForClear();
Me.BuzzHorn(Max_Vol);
}
I thought true freedom came only when you had nothing to tie you down?
Control is an illusion, order our comforting lie. From chaos, through chaos, into chaos we fly
... suicide car bombers.
Oh, let's see - maybe the fact that I have to share the roads with dangerous drivers?
We limit the rights of some to protect the rights of all - if you are an unsafe driver, I will happily limit your right to drive if it increases the rights of the majority to drive safely.
That, my snide friend, is what gives me the right - the same right that pretty much all of the laws of the US are based on. Also the same reason you have to take a driving test and maintain a driver's license. Yes, that's right, a license to drive. Pretty "Soviet", eh? In your view, is it only American if we just let everyone jump behind the wheel, even the blind and insane, because "America, Fuck Yeah!"?
I'm sorry, but think before you post. It enriches us all.
I almost laughed out loud when I saw pictures of MIT's pimped out Land Rover. Besides the numerous external sensors and other gear mounted on the vehicle, I read that there is so much internal equipment to manage everything that they had real heating issues that were solved by installing an additional air conditioner and a power generator to power the AC. This is what happens when you give some money and parts to a bunch of bright geeks with too much time.
Here.
--
WHO ATE MY BREAKFAST PANTS?
Most freedoms are privileges (from a pure survival standpoint I mean), yet we've made them rights because we feel they make for a better society overall. Be VERY careful whenever you want to clamp down on something we've had choice in for quite a long time.
I would be interested to know exactly what the scores where and how they were derived. From reading the Popular Mechanics (blog) and Register (blog) reports, it sounds like Stanford might have gotten a bit of the raw end of the stick.
Specifically, the Register is reporting that it DARPA counted the up to 20min Stanford's car was stuck sandwiched between two other cars due to Cornell's robot screwing up against it, and Popular Mechanics is reporting that DARPA says Stanford lost to Carnegie Mellon by about 20min.
Sounds like it would have been a extremely close race if DARPA had been applying more reasonable (from the principle of trying to eliminate luck as a large factor) accounting principles.
Firstly, a car you can't ever drive would never sell in the US. People want control, they want the ability to drive off-road even if they never actually do (see SUVs), and they love their older cars too much to stop driving them. Secondly, even if every car was automated, that would only take care of a *few* of the problems faced by automated vehicles. They would still have to deal with all of the problems that are caused by things other than unpredictable drivers, such as: wind, rain, snow, ice, fog, loss of GPS, worn or obscured road markings, people walking in the road, things that fall from trucks on the freeway, tires that blow out, malfunctioning traffic signals, downed power lines, mechanical failures of all kinds, collapsed bridges, avalanches, sinkholes, people trying to trick the robot sensors, and all the other problems I didn't happen to think of just now. If you really want your robotic car to be 100% safe, you have to program it to handle so many varied situations that I believe programming it with traffic rules for safe driving around humans would be a relatively small part of your work.
Now it's true that you could drive more efficiently without humans, but that will have to be phased in gradually. For example, you could have special robot lanes, and perhaps eventually entire robot-only streets in big cities. But that would only be possible *after* the introduction of autonomous vehicles.
main(c,r){for(r=32;r;) printf(++c>31?c=!r--,"\n":c<r?" ":~c&r?" `":" #");}
You're exaggerating, in the extreme.
I'm willing to bet every (human) driver in this country would have succeeded with flying colors on this course as well. In fact the odds of a driver getting killed in an accident any specific day are extremely slim, and they'd be much smaller still, if you restrict that to low-speed driving, during the day, etc., etc.
The skill of these robotic drivers can only be determined with any reasonable accuracy after they have driven many MILLIONS of miles. Only then can you say they are, on average, safer than human drivers. And even then, it would still be insanely ridiculous to claim they drive flawlessly.
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
Transition is the key issue. If we were to redesign the transportation system again given the current state of knowledge and technology, it would probably be vastly different than the system that is currently in place. However, there already is a system in place which is crucial for every aspect of our lives. So a feasible transition plan will have to be central in any new technology.
Similar tech is in use daily on large farms, but it is a rigidly defined route. Self steering tractors are very common now, there are even kits you can get that bolt on to your normal tractor. They are more intended for keeping precise plowing/tilling/planting etc spacing, where inches count highly, but using GPS and maps of the fields they work perfectly fine. If there was a dedicated lane next to existing freeways for slow and steady cargo delivery-separating human drivers from the bot drivers- this could be done today fairly easily I think using similar off the shelf stuff. The darpa challenge is way more about building autonomous robotic fighting vehicles/ military convoy vehicles (Read the oshkosh terramax site, why they are using that large truck in their efforts, it is a direct sales model if they can get it to work right, as they didn't this test), and as such needs to be loads more complicated than just following a wide and clear road with traffic all flowing the same direction, etc. One of the larger problems is off the wall events that can't be adequately programmed for in advance and have to rely on sensors, like the random deer out in the road, people running across the highway, "road gators" and other unexpected trash in the road, stuff like that. In a military situation, perhaps they wouldn't bother, smash their way through, but still try to not run over all the locals during the trip.
I think it is going to be really hard to come up with a civilian model that would work on all roads, just too many variables to contend with. In a war situation they can afford to be a little more sloppy in the collateral damage department (from their point of view, not the other guys of course). They want to pull expensive humans out of the mix as much as possible, while still retaining near the same level functionality. On a civilian road during non war conditions, the quality of the self steering needs to be loads better.