Slashdot Mirror
CMU First To Qualify For DARPA Grand Challenge
Anonymous Coward writes "As of 18:00 March 9th, Carnegie Mellon's Red Team is the only entry to successfully complete DARPA's Grand Challenge Qualification Inspection and Demonstration (QID) before the main event on March 13th. The NY Times has this article detailing this first step towards winning the Grand Challenge."
No reg needed!
Join the Free Software Foundation
One of the other competitors is from my university. Looking at the relative sizes, I hope the hummer in the article doesn't get in TerraMax's way.
Prospective station wagon buyer: "I know what you say is true...but...er...I don't know how to maintain a tank!"
I can't be the only one who questions motives when the $1M prize is being sought after by a team with more than $2M already invested. What is the eventual payoff?
Linux: Free if your time is worthless.
Cars like this will fork into commercial/environmentally friendly vehicles and on the other side MadMax like racing cars!.
I Still cannot understand why Bush hasnt pushed SUV fuel consumption to better levels, you guys probably have the technology to make cars run on water.
"The quality of life is inversely proportional to the number of keys on your keyring."
Just out of curiosity, how well would the rovers' "route picking" routines cope with this challenge?
;)
I read that the operator says "go from here to here" and the onboard 'AI' chooses the best route in a 3d visualisation - is this software open-source, and could it be used in this challenge? I can't see any major differences, other than the relative lack of parked cars on Mars (2 pathfinders and a beagle, iirc)
Should DARPA have emailed NASA before starting this?
The Slashdot Paradox: "100% Overrated"
The real prizes:
the knowledge gained throughout the project
getting one's name published for taking an active role in the project (which can lead to further opportunities)
the overall experience, i.e. 'Hey, I did that"
The pursuit of intellectual challenge is not about money...
I want to drag this out as long as possible. Bring me my protractor.
they divided your time by the cost of your machine.
its impressive when you build a mega$ robot, but a minimal robot that manages to finish is way cooler
TAIWWP :(
Does anyone know of anywhere hosting pictures of these unmanned robot vehicles?
... doesn't this basically lead directly to the US military dropping off Robo-Tanks in foreign countries as they please? We know that a steady diet of wars figure heavily in the plan for the forseeable future. The Robo-Tank cuts down on friendly casualties, thus making conflicts more palatable to the public.
Now I find this as cool as anyone else, from a technological standpoint. And it definitely has civilian applicability. But let's face it, this contest isn't about finding cheaper ways to haul cargo or reach remote locations.
I would sure hate to be a geologist, prospector, or hermit in the desert that day.
Gelogist: [mumbling to himself] Finally! Proof that the formation of this arroyo was caused by--
[Geologist is flattened by an army of driverless cars driving at upwards of 60 mph, one of which detects the collision too late and actually backs up, running over him again, as failing avoidance mechanisms kick in]
Any generalization is a stupid one.
Seeing as DARPA wants to turn this technology into a military robotic transport, I don't know how valuable it's going to be if it has to be pre-programmed with terabytes of data just to move. What about if they invade somewhere they don't have good maps of? Somewhere with a dynamic landscape (desert, rocks etc)?
I'm all for innovation, but exploiting poorly-worded rules just to win for winning's sake is an empty victory at best.
The Qualification Inspection and Demonstration was rumored to be the state's Driver's License Test.
The vehicles had been fretting about the dreaded parallel parking portion of the test.
Nice to know 3 million dollars buys you a roll over and placement. What the hell did they do to the HUMVEE that made it roll over? I know HUMVEEs and HUMMERS, they don't roll easily.
I'm surprised ay the number of Cannonball Run references and the low number of Death Race 2000 references
The one where supposedly as long as the correct letters are there, and not necessarily in the correct order, the words can still be understood?
Am I the only one who at first glance thought this was a story about pron and Prada?
Don't park drunk, accidents cause people.
All competitors are given the actual route as a series of GPS waypoints a few hours prior to the race. Red Team is going to send those waypoints back to CMU, have the big iron there figure out the best course based on all the map data, and then download that course to the robot prior to the start. In a way this is cool, but it seems like they are using a loophole. A much more interesting problem would be to navigate a course that you know nothing about other than the waypoints.
The other teams are using techniques that require more onboard intelligence and route finding. The most interesting vehicle is from Cal. They have a motorcycle. Even though I went to Stanford I am rooting for the Cal motorcycle to do well since they have the most unique vehicle. Hopefully the team of Stanford alums (already dropped out) can come back next year and beat them.
Lasers Controlled Games!
Weird title, seeing how they showcase the CMU entry, a high-school entry running in an Acura (donated by a parent who works for Honda) and a single-member "team" trying to do a motorcycle entry.
Here's a size comparison from the Oshkosk website... their truck is 9 feet tall, a hummer is 6 feet.
HIV Crosses Species Barrier... into Muppets
Nice photo of the vehicle. But if you know Umit, tell him to get his photo retaken! It's difficult enough to do a good shot in a striped shirt, but those are wrinkles!
Seriously, I'm a photographer- tell him to get a new press head shot.
Pretty slick design tho. I somehow think this is going turn out to be a contest of overdesigned rather than 'clever'....
"A truly autonomous vehicle will revolutionize land transportation. It will be a new tool certain to improve mankind's quality of life.." Red Whittaker Team director
..
Wrong priority, what about revolutionizing land transportation with a non polluting vehicle
I feel kinda priveleged to be part of this. DARPA is working with SCORE International Off-Road Racing (http://www.score-international.com) to do the "checkpoints" and road crossings for this event, of which I am a part of.
Basically it means sitting around all day waiting for these things to show up, but it will be fun nonetheless.
I have a feeling that this event will not have a finisher, but from what I have heard DARPA plans on carrying out this challenge for about five years anyway.
One of my friends works with them (I'm a student at cmu), and recently he told me that they flipped the thing and crushed a bunch of shit. Lots of the roof-mounted equipment had to be replaced. Apparently it took a turn too fast. I'm glad they were able to have it ready, and a bit surprised they were the first. They certainly seem to be striving to do their best. Anyway, knowing the red team's capabilities, we won't see that problem again. Here's to hoping and their success.
I'm somewhat surprised that so many young people would work on a project that will help our military develop unmanned hunter-killer vehicles. Isn't this why DARPA's funding this project? I'm not against the project, just curious if there are any conscience issues involved here.
Speaking of automatic tanks, did any of the teams name their entry Ogre?
with this is simply how cheap the US military is getting away with this. Instead of setting forth a proposal, taking bids, working in tandem with one of the big development houses, they offer up a rediculously small prize. If they had gone through someone like Lockheed Martin, they project would have easily cost them into the 100 million dollar range. Oh well. Hopefully the military will get what they paid for.
Yup, those are some pretty cool prizes. But we gotta remember other prizes, like bragging rights. I'm already sending this article all over to my fellow CMU almuni friends, to other non-geek friends, etc. All this, and I have nothing to do with robotics and graduated almost ten years ago. Woohoo! I love bragging rights...
"He uses statistics as a drunken man uses lampposts...for support rather than illumination." - Andrew Lang
Anyone know if there are plans to televize or broadcast the race in any form?
I can't say for sure, but it seams to be that with all this equipment on roof, vehicle's center of mass is now much higher than it used to be, making the vehicle much more instable.
No sig today.
Immediately springing to mind is the scene from 'Ocean's Eleven' when the Monster truck is racing the R/C mini-monster truck, and it runs it over.
Have some respect, please. That scene is from the new, and not as good version of Ocean's 11, not the original.
Notice:
DARPA has subsequently changed the event from "Grand Challenge" to "Most Extreme Elimination Challenge" in hopes of drawing more support and viewer coverage. Fox is asking for broadcasting rights.
SELECT * FROM USERS WHERE A_WINNER = "YUO";
You are right that it would be much cooler if the robot could find its way without the map, but for the purposes of the military it is plenty good that it reads a map and only does obstacle avoidance in real time. This is the way that a tomahawk cruse missile works and it works pretty well. One of the first things that the military does in a conflict is update the maps and reconnaissance for the area.
This also would great as a civilian technology. If I could drive my car to the interstate and hit an autopilot button so the car would follow a preplanned course and alert me if it saw a problem - that would rock. I wouldn't care if it only ran the speed limit if I could read a book and the car handled the bulk of the driving.
I have great faith in fools - self confidence my friends call it. - Edgar Allan Poe
Attendance was about the same today except it didn't appear that there was as many media representatives present. Again temperatures were in the 90's. I acquired a media pass today and was allowed access to almost every area of the speedway including the pits and the start line. This will allow me to film each entry up close and interview members of the teams. DARPA is also publishing the daily events here and here.
Vehicle inspections on the rest of the field were performed today. DARPA is reacting as fast as they can to modify the rules and give every opportunity to each of the teams in hopes they will be able to qualify. DARPA is now allowing the teams as many appearances on the Q&D course as requested by the teams. The Q&D that was scheduled today became an opportunity for teams to iron out their problems on the track.
The Blue team with the CyberRider (the motorcycle entry Web Site) was the first on the Q&D course. It traveled about 20 feet when it fell over and exposed it's greasy side.
Team ENSCO Web Site traveled to the first major turn and failed to navigate the first sharp turn. This is also the same place Team TerraMax web site failed today and the Sci Autonics web site team.
The first turn appears to be difficult for the vehicles that make it there.
The ASI/Florida State Web Site had two more runs today. They disabled all of the perception systems and successfully ran about 1/3 of the course. The first run, the vehicle was driving like a "drunk sailor" according to a member of the team. Florida State took the recoded path data to tune their vehicle controller. ASI reported the vehicle tracked much better on the second run. I get the impression that a few more Q&D course test runs will be performed before the vehicle is ready to qualify.
Team Caltech Web Site made two more runs on the Q&D course. Well, both runs were consistent, but not as good as yesterday. Both times, cleared the start line and banked hard left as if it were going to the last way point. I hope to find out today what happened.
There were several other teams that attempted the Q&D course today with similar or worse results.
The highlight of the day was CMU's Red Team web site. As anticipated they made it to the finish line of the course and electrified the spectators and increased pressure to the rest of the field.
That's the highlights of the second day of Q&D testing.
For the extra paranoid (or visionary, depending on your leanings,) check out:
Robitic Nation Evidence run by Marshal Brain, of "How Stuff Works" fame.
He envisions a future where autonomus military vehicles could be ordered in to action against any foe, including California.
You might be interested in knowing that, according to AW&ST, the army/air force in Iraq has found in many cases that it is more efficient to transport cargo within Iraq via C-5, instead of 12 large trucks.
The reason was because the loading and unloading areas could be secured but not the highways in between.
Check out the February 23rd Issue.
myke
Mimetics Inc. Twitter
I suppose that the Internet (former DARPANet) is some insidious hunter killer commmunications device?
I run one of the Grand Challenge teams, Team Overbot. We have a vehicle (a modified six wheel drive Polaris Ranger), a shop in Redwood City, funding, equipment, and people. We're well along; the vehicle has most of its actuators and some of the sensors working, and about a third of the software is running. We're one of the five DARPA-accepted teams.
Many of us are Stanford alumni or students, but this is not a Stanford project.
Our basic technical approach is to build a rugged, reliable vehicle with conservative control strategies. Others may be faster, but we expect they'll get into trouble at high speed. Our top speed is 40MPH. The real problem with the Grand Challenge is not going fast on the easy parts; it's getting through the hard parts.
The 6WD chassis we're using is one of the most bump-tolerant platforms around. It can go over railroad ties at top speed without problems and without going airborne. The center of gravity is low. The front and mid axles have independent suspension; the rear axle is a swing arm. This simplifies low-level vehicle control. All wheels can be driven, although at higher speeds, we will switch from 6WD to 4WD.
We have five computers on board. Three are small PC/104 machines, and two are Pentium 4 machines. All run QNX (the OS for when it has to work.) All are industrial-strength ruggedized units. The actuators are all servomotors driven by industrial microcontrollers. All this hardware is off-the-shelf industrial control gear.
Sensors include LIDAR, doppler RADAR, sonars, cameras, INS, GPS, etc. Some of them are used in unusual ways. That's all I'll say about that.
The pathfinding strategy is indeed borrowed from video game technology. It's more structured than Brooks-type behavior based robotics, and it's less structured than Latoumbe-type planning. There are three layers of control; the top one we call the "back seat driver", because it has only advisory authority over the "driver".
We have road map and topo data onboard, but it's used more as a hint than as rigid guidance. We take the waypoints DARPA gives us (on a CD, at 0430 hrs the morning of the race) and load it in. There's no offline preplanning. Wouldn't help in the real world.
If nobody wins this year, which is quite likely, we'll be back next year with a faster vehicle.
Post questions and I'll answer them here.
John Homogle
Team Overbot
Hopefully, this is just the first step towards that goal.
Given the 40 or so years between the DoD's ARPA net experiments and what we have now, it isn't that far fetched.
The point of the Darpa project is to advance technology for driverless military vehicles, primarily for convoy work. To my mind, creating a computer system to quickly plan out routes based on intelligence is an important part of a practical solution.
Not only does it more accurately reflect the technology's intended use-case in the military field (convoy operators would lilely be given a general route a couple hours before a mission, instead of simply told, 'get it to this point and leave right now') but it also means that more of the technology is outside the vehicle.
A cost-effective solution would need to have as cheap a vehicle as possible. While a fully autonomous system might be nice for a science fiction 'technology run amok' film, in reality it's more effective to have sparse mobile systems with an ops center capable of planning routes for several vehicles.
It also costs less when one goes 'wheels up' or is captured by the enemy.
Kevin Fox
I could have saved them $250,000 in sensors by installing a $100 roll bar!
The only thing that is yours, is your soul; everything else is borrowed.
haha...I work at the hosting company that host the redteamracing.org website, and yes we are getting slashdotted to a certain degree.
If the links to the movies are slow or time out please try again later as we are working with them to setup mirrors.
thanks!
itall
DNT(DO NOT TEST)
2 years and no mod points. Join reddit. Because openness is good.
Sandstorm had been driving with new code for reducing speed in the corners.
Slashdot user Animats is John NAGLE and he obviously posted the original.
No time to look for the link.
Repost trolls are getting out of hand.
I went by the event yesterday as a spectator and got to see Red Team do their run. Of the 23 teams who made it this far, they're the only one that has completed the qualification course so far. People complain that they have a more accurate map and that they're not doing real AI, but based on their performance on this surprise course, they have a real obstacle avoidance system.
In one section there was a minivan parked in the center of the GPS path. Of the eight vehicles I saw run, only three made it past the car. Three hit it, and the rest failed before making it that far.
It seemed that the biggest problems teams had were getting GPS right. Several drifted off course or turned the wrong way, going off course. One got the next GPS coord inside of its turning radius so it kept circling a spot until they turned it off.
Lots of great designs though, and some really impressive engineering.
You can look at the latest pictures, the teams' relative positions, and status charts.
http://www.grandchallenge.org/
washingtonpost.com spent three months following a Northern Virginia team as it conceived and built its DARPA Grand Challenge vehicle. Check out the video, and read a related story from today's Washington Post.
I withheld midterm grades of our course, pending the outcome of QID, since this is a third of the grade. Since QID went well, grades will be good, and it will be a pleasure to assign these.
Withold the grades, and students will do anything. (Note to Theory profs out there: withhold grades until someone proves P != NP )
so how does an ethical collision-avoidance system deal with human hacking of the route??
eg. a huge number of people just go stand in the way of the robotic transporters - like, making lines across the routes
how do the transporters figure out what to do? do they just run these people over? or does their collision avoidance system make them turn around and come back the way they came....?
Put the prize on a sliding scale so that achieving the distance at a slower speed would still have its rewards. They could then use the working method(s)to build up to the speeds they want. Rather than trying to do it all in one miraculous leap.
You can legislate morally you can't legislate morality
Is it too late for me to enter my Roomba?
Towards the end of this year, PCI-Express and next generation graphics cards using them will come to market.
How is this significant for the DARPA race?
Well, newer generation graphics cards with highly programmable graphics pipelines can act as very powerful SIMD processors. Up until now, their capability has not seen much use outside of graphics because the AGP bus allows data to travel at full bandwidth in only one direction at a time. This meant that every time you need to download data from the video card, you had to flush the AGP bus, loosing or delaying the uploaded data.
With PCI-Express, data could travel both ways concurrently at full bandwidth, so there's the potential for using the graphics card as a specialized SIMD processor.
I bet much of the processing for the vision and obstacle avoidance could be done on a GPU. If that's the case, instead of having a 1.5 gigaflops CPU per pc, you could have 10 or 20 gigaflops (IIRC) of processing power at your disposal for little over $1000, thus making the necessary computing hardware much cheaper.
Currently, the CMU team, for example, has multi-itanium servers aboard their Hummer, which is NOT something doable on a shoe string budget.
The Blue team with the CyberRider (the motorcycle entry)
Ummm... the Blue Team's motorcycle entry is known as Ghostrider. CyberRider uses a 4-wheeled vehicle.
Be faithful to your obsessions. Identify them and be faithful to them, let them guide you like a sleepwalker. JG Ballard
Man, these teams are using the wrong military vehicles. Humvees? Supply trucks? What the hell is that? Seriously man, strap a GPS system onto an M1 Abrams and open that baby up.
I say screw collision avoidance. Go for collision dominance. Any obstacle capable of stopping a 65 ton tank travelling at 45mph is gonna show up on the mother f**king map.
Here are some photos:
DARPA Grand Challenge Photos
Enjoy!
Don't believe me? Stand out there in the field while I revv the engine...
Debunking the "59 Deceits"
All competitors are given the actual route as a series of GPS waypoints a few hours prior to the race. Red Team is going to send those waypoints back to CMU, have the big iron there figure out the best course based on all the map data, and then download that course to the robot prior to the start. In a way this is cool, but it seems like they are using a loophole. A much more interesting problem would be to navigate a course that you know nothing about other than the waypoints.
To misquote 'Red' from an article, some people have accused his creations of not being robots because they do one thing, and one thing well with out to much intelligence. This was his plan from the begining, get from point A to point B without crashing. It doesn't need to be intelligent, just elegant.
I can't quite parse your first sentence there, but I think that you are saying that I am saying it isn't a robot. I have said no such thing. Read my other posts. I think it is a valid solution, and it looks like they will do very well. Hopefully it won't flip on the open road again.
Lasers Controlled Games!
You tried to oppose a networked military.
As the Iraqi Army found out, the Internet kills very very well.
I would be a bit surprised if they decided to randomly do S-turns at 35 mph on open road during the competition, so I would guess that won't be a problem. As for situations where the course might require them to do an S-turn, I imagine they'll have tweaked the programming to slow down from 35 mph before doing so in the future.
"They" don't decide anything once the competition has started, but I know what you mean. The website makes it clear that turns will be slower. It seems odd that they did do an S-turn at 35 mph during testing right before the comp though.
Lasers Controlled Games!
Unless you are dumb enough to believe that the 1 meter granularity of the maps is enough to drive. It isn't. And by the way the red team is a continuation of the NavLab project that has had a vehical capable of driving itself for about 10 years now. It is getting better each year. Now it, w/o a map can drive over a dirt path or a crowded highway. The issue here is speed. If I already know the rough path to follow I can drive faster because smaller corrections and fewer surprises will await the vision systems that actually drive the vehical. ITS A RACE DUMBASS. The idea is to go fast AND finish the race. The CMU team is good enough where just finishing isn't enough anymore. Nobody else is at that level yet.
Also, there isn't one vision system but five that vote on the correct course of action, in additon to the radar and the lasar avoidance detection. Finally, if you think the CMU team will get beat by a Standford team, I've got a bridge to sell you. Do CMU robot teams lose? Sometimes (not very often) but when they lose, they are beat by large corporations generally from Japan that put large amounts of resources into the project. Other teams from other schools GT, Cornell, MIT, generally will give CMU a run for its money but rarely actually beat the CMU teams. For more info, check out the robocup. CMU has won more than ever other organization in the world combined.
Good Luck, sounds like you'll need it.
On the Chinese front, please accept an American apology for the implication of the above post. (And domestic surveilance... GBA..) ...we're not intent on attacking you, even if our relations aren't perfect. At least, I don't think so.. but after Iraq, who knows? I'm voting for Kerry... I'm hoping we won't have to worry about that.
http://www.grandchallenge.org Run by DARPA.
A virtual view of the DARPA Grand Challenge.
Live Tracking will show relative positions of the Challenge entrants, and requires a 7 MB download each time you use your browser to view the tracking. (Although this is still under construction apparently)
The Status Board provides a 30 second update of the status of each Challenge team.
The Image Gallery will contain the most recent images from the Challenge, updated nightly through March 14.
GM bought the Hummer name from AM General to cash in on the reputation the Hummer has built. The H2 has NOTHING in common with the original Hummer except for the name.
H2 is built on a Tahoe/Suburban platform (not Tacoma, that's a Toyota). As for its off-road prowess, just look at its front and see how little real ground clearance there is. IMHO reviewers were influenced by the fact that this oddity is called Hummer and might have assumed it inherits anything from the original.
I will pay $100 for the name and address of the person responsible for that posting.
John Nagle
nagle@overbot.com
Mark my words, unarmed civilians will be killed by American war bots.
'Course the excuse will be that they were harbering terrorists, or that there were armed terrorists in the crowd, or that terrorists were formally at that position so it isn't our fault and we all regret the loss of innocent lives but this is war so death happens so shut up and watch Survivor XXXIV but us guilty of war crimes? NEVER!
Debunking the "59 Deceits"
Here is the link to DARPA's coverage on the events: results
DARPA is continuing to lower the requirements to qualify for the race in hopes to allow more than about 5 contestants. They are also overlooking some of the safety rules in the pit area to allow the teams to do as much troublshooting as possible without a DARPA rep. being present. Originally teams were not allowed to move their vehicle without a rep. present or even put it into autonomous mode. The teams are now being allowed to drive their vehicles in manual mode and park them in autonomus mode in an attempt by DARPA to allow as many teams to qualify as possible.
The big news today is that three more teams have successfully finished the Q&D course.
Team Caltech. The Caltech team is the first to completely navigate the course using a NavCom GPS receiver.
Sci-AutonicsII is the first to unequivocally clinch a position in the Grand Challenge competition by successfully navigating the Q&D course two times. Sci-AutonicsII has decided to pass on their originally scheduled second run of the Q&D course.
Virginia Tech is the fourth team to completely navigate the course.
Team CIMAR (ASI & Florida State) made two more runs on the Q&D course. Their second one was made with reflexive obstacle avoidance functioning but without IMU integration. They were on their way to complete the course until they passed under a walkway that briefly blocked GPS. The dropout caused the vehicle to veer hard right and swap paint with the retaining wall. The damage was only superficial despite brutally applied custom pinstipes to one of the Preco RADAR sensors.
I find it interesting that no one pointed out that Carnegie Mellon smashed into the final obstacle in its first QID run. The vehicle didn't seem to sustain any damage, it was just rather surprising/funny... especially in light of the Pittsburgh Post news article quoting a CMU student, "Sandstorm stopped on its own... just like it was supposed to." (http://www.post-gazette.com/pg/04070/283656.stm)