Stanford runs quad core for Urban DARPA challenge

Author: Wil Harris
Published: 17th February 2007
Comments (15) Stumble
The Intel Stanford Urban vehicle.

The Intel Stanford Urban vehicle.

Today, the Stanford University DARPA team is announcing that its entry car for the 2007 Urban DARPA challenge is powered by low-power Core 2 Quad chips.

The DARPA Challenge, we're sure you'll know, is the challenge to get an unmanned vehicle across a course in the desert. The Urban challenge is a variation, taking place in a city environment.

The Stanford car is running Core 2 Duo and Quad processors, and this is notable for one reason - this year, it's the only car that doesn't require a secondary power source. Since the Core 2 processors are so low power, they can run off the car's alternator - a big advantage.

David Orenstein, one of Standford's engineering chaps, wrote thusly:

"“In the last Grand Challenge, it didn’t really matter whether an obstacle was a rock or a bush because either way you’d just drive around it,” says Sebastian Thrun, an associate professor of computer science and electrical engineering. “The current challenge is to move from just sensing the environment to understanding the environment.”

That’s because in the Urban Challenge, sponsored by the Defense Advanced Research Projects Agency (DARPA), the competing robots will have to accomplish missions in a simulated city environment, which includes the traffic of the other robots and traffic laws. This means that on race day, Nov. 2, the robots not only will have to avoid collisions, but also they will have to master concepts that befuddle many humans, such as right of way."

The finish line is a way off, since the competition doesn't start until November. But you can bet that the Intel and Stanford guys are going to be working pretty hard until then to make this work.

Let us know your thoughts on the DARPA Urban Challenge over in the forums.
Quote DeX 17th February 2007, 16:57
The Stanford team did a great job in the last DARPA challenge but I reckon this one will be as disappointing as the first desert based race. No car will get to the finish. There's no way they can program for even a small subset of all the possible obstacles and random events they will come across. All the same, I can't wait to see how each team does! ;)
Quote randosome 17th February 2007, 18:28
wow they had to use other power sources before ? i didn't know they would have been drawing so much current
Best of luck to them though, itl be great for someone to actually complete this challenge one day
Quote DXR_13KE 17th February 2007, 18:42
destruction derby anyone ? :p
Quote Kipman725 17th February 2007, 19:22
Quote:
Originally Posted by randosome
wow they had to use other power sources before ? i didn't know they would have been drawing so much current
Best of luck to them though, itl be great for someone to actually complete this challenge one day


last time they all had to use things like beefed up altanators. Need quite alot of proccesing power to navigate efectivly aparently.
Quote rowin4kicks 17th February 2007, 19:32
wow that is cool, and i bet intel is beefing in some monies to that project!
Quote EQC 17th February 2007, 19:50
"the competing robots will have to accomplish missions in a simulated city environment, which includes the traffic of the other robots and traffic laws."

I wonder how they'll deal with that...what if Robot B rear-ends Robot A, and both become disabled? It'd be a shame to "lose" the competition because some other team's lousy car crashed into your good one from behind.
Quote TheColdLord 17th February 2007, 22:49
Interesting, t'would be interesting to see if, after thne race and if anyone will finish, will there be any other implementation of the technology.
Quote bubsterboo 18th February 2007, 00:09
I hear PC cars crash a lot.
Quote Snafu-X- 18th February 2007, 01:54
Are bad puns a bannable offense? No? Maybe they should be.... :(
Quote crazydeep74 18th February 2007, 03:40
Quote:
Originally Posted by EQC

I wonder how they'll deal with that...what if Robot B rear-ends Robot A, and both become disabled? It'd be a shame to "lose" the competition because some other team's lousy car crashed into your good one from behind.
I wonder which computers insurance pays the damages.....
Quote Dezza 18th February 2007, 17:31
Quote:
Originally Posted by EQC
"the competing robots will have to accomplish missions in a simulated city environment, which includes the traffic of the other robots and traffic laws."

I wonder how they'll deal with that...what if Robot B rear-ends Robot A, and both become disabled? It'd be a shame to "lose" the competition because some other team's lousy car crashed into your good one from behind.

We could see the first instance of Robo Road Rage
Quote supermonkey 19th February 2007, 00:44
Quote:
Originally Posted by EQC
"the competing robots will have to accomplish missions in a simulated city environment, which includes the traffic of the other robots and traffic laws."

I wonder how they'll deal with that...what if Robot B rear-ends Robot A, and both become disabled? It'd be a shame to "lose" the competition because some other team's lousy car crashed into your good one from behind.
That presents one of the design challenges. I can understand why the Defense Department would be interested in an automaton that can drive defensively, as well as offensively.

-monkey
Quote K.I.T.T. 19th February 2007, 09:54
Quote:
Originally Posted by Kipman725
last time they all had to use things like beefed up altanators. Need quite alot of proccesing power to navigate efectivly aparently.

[analy retentive mode]

...well when your doing up to 60MPH and needing to map the input from 3 pairs of laser range finders into a realtime constantly changing 3D model then decide what part is flat/driveable, what speed to drive at then overlay that (in the case of Stanley) with a video feed from a camera to decide where in the long run you can drive before taking in to account the GPS way points they were given and actually doing anything about it i would think that takes a fair bit of processing power....just a bit of course.

they were using backup generators for some of them (like the two Carnegie Mellon hummers)...probably best idea as well save the engines power for driving and running the all important AC to keep all the equipment cool

EDIT: one of Carnegie Mellon robots (Highlander IIRC) had 7 Pentium M systems to do the processing, so the kind of power you need is that and a bit more

[/analy retentive mode]


Its going to be so cool to see how they handle the new urban enviroment, basic navigation should be easier but you have to take so many more things into consideration, i'd love to be able ot do something like this! :D
Quote otispunkmeyer 19th February 2007, 15:53
Quote:
Originally Posted by EQC
"the competing robots will have to accomplish missions in a simulated city environment, which includes the traffic of the other robots and traffic laws."

I wonder how they'll deal with that...what if Robot B rear-ends Robot A, and both become disabled?

they can compute who's fault it was on the spot and file lawsuits accordingly via wifi.
Quote Redbeaver 19th February 2007, 20:06
Quote:
Originally Posted by otispunkmeyer
they can compute who's fault it was on the spot and file lawsuits accordingly via wifi.

i love this comment...


but yeah, some quadcore should provide plenty of overhead incase the car needs any extra power... i say more headaches will go into the programming on adequately balance the processing power to the AI... whoever can program the most efffective AI will definitely reach the furthest.

and yes, id put my money on nobodys gonna finish like the last DARPA lol.

i made an AI mouse robot before that navigates a maze, and thats headache enuf lol
Log in

You are not logged in, please login with your forum account below. If you don't already have an account please register to start contributing.





Stats: 0.045 seconds