Meostro writes "DARPA announced the 3rd "Grand Challenge" today, The DARPA Urban Challenge. "To
succeed, vehicles must autonomously obey traffic laws while merging into moving traffic,
navigating traffic circles, negotiating busy intersections and avoiding obstacles." This year's new twist is two tracks for entry: the first is the same as the previous two challenges (develop on your own without Gov't. funding), but the second involves "submitting a detailed proposal for up to $1 million of technology development funds." Here is the PDF press release ."
As a tax paying citizen of the United States, it sure is frustrating to try to find the results of DARPA research.
Yes, they do research in defense but shouldn't there be a little more than a tiny graphic or blurb about what work they're doing? Couldn't they at least take the time to write an abstract or 1-2 page paper with unclassified information on each project?
Instead, I find the following links in the 'Archives' [darpa.mil]:
My alma mater has produced better papers than this in these fields. I know that a lot of this stuff isn't classified and they list their programs on their sites, why can't they do a better job in showing the American public what they've done with our money?
The Grand Challenge Forums [darpa.mil] are flooded with only vendors. Where are the designs and reports by the teams from older Grand Challenges? Why isn't this structured more like RoboCup where the learning algorithms are released every year so that future contestants can build on this?
The fact that this contest isn't run in a more open way makes it seems like less of a "contest" and more of a "do our research for us!" kind of thing.
It is likely that DARPA has become incapable of inovation because of internal politics so they need to attract new ideas from the out side.
DARPA is a funding agency, not a research institution. They actually have very few employees, who are mostly there to identify promising research areas and allocate money to invest in them. The actual research is done by academic and industrial research groups. Incidentally, many if not all of the DARPA project managers are actually very good research scientists and engineers who take 1-3 years away from their normal work to work for DARPA, not career bureaucrats (wow, that's a hard word to spell) mired in politics. Most of them really view it as an important public service.
You'd rather find out DARPA has handed out a $10B contract to a "regular" defense contractor to solve this particular problem in ten years, only to find out at year #9, they will need a seven year extension and have cost overruns of another $8B?
A couple of years, a few million dollars as one carrot, the other two are establishing legitimacy in a captitalistic market looking for established technology, and finally, chest thumping in the geek world is a very, very tiny investment by comparison.
As a tax paying citizen of the United States, it sure is frustrating to try to find the results of DARPA research.
Yes, they do research in defense but shouldn't there be a little more than a tiny graphic or blurb about what work they're doing? Couldn't they at least take the time to write an abstract or 1-2 page paper with unclassified information on each project?
Try instead going to Google Scholar or another academic index and searching on the titles for various DARPA projects. Having worked on several DARPA-funded projects, I can tell you that there is generally a significant emphasis on publishing results. DARPA-sponsored work probably results in dozens or hundreds of articles in scientific journals a year, all of which are available to the public.
When you say your alma mater "has produced better papers in these fields" you should have a look at the acknowledgements section of these papers. Chances are pretty good many of them will have a statement like "This work funded in part by DARPA (or NSF, etc) grant number XXX."
"The fact that this contest isn't run in a more open way makes it seems like less of a "contest" and more of a "do our research for us!" kind of thing." Well, yes.
This ISN'T about open technologies, this is about building up to a defense (warmaking) capability.
On the one hand, you have to release a tiny bit of information, just so all the competitors have the same basic assumptions and stay in the same universe of solutions. On the other hand, you don't want to constrain the competitors too much with regard
...how this can be "safe" (as far as it can be anyway) with cars which are automated going on roads for which the system works (I'm assuming most are based on the idea of staying within the white lines) I worry about this quote...
We believe the robotics community is ready to tackle vehicle operation inside city limits.
- Dr. Tony Tether, DARPA Director
You can build the safest car in the world but there is always a need to be able to take a very quick decision to avoid some other idiot who might be breaking the rules of the road and not be in an automated car... still, if we all had them...
There are an awful lot of streets that do not have white lines, or have cars parked on both sides so that there is really only 1 real lane for traffic even though its a 2 way street. What does it do when it reaches an obstacle that avoiding it requires you to break the rules?
You would needs some serious AI and pattern recognition to really replace drivers. There is just too much that can go wrong.
I think that is the point... Driving like chess has very basic rules that are easy to learn and difficult to master and on top of that you have the complicaton of other actors (other drivers) mixed in. In the end if you create an AI that can preform a task as complicated as driving that same AI should be able to learn to do just about anything
So you expect a chess playing computer to play calvin ball (ie rule changes spontaniously in the game) Most AI can't handle this, as most self driving cars in the future still won't be able to handle these things. What you will have are cars that self drive only on well marked roads, perferably with GPS indicators of what these roads are. You will be forced to manually drive small narrow roads.
Well my point is not that these AIs will play "Calvin Ball" but in developing driving AI that can deal with more and more unknowns or independant actors will lead to advances in computer science and robotics.
Are you saying that we shouldn't pursue the study of self driving cars simply because you believe it's a tough cookie to crack? Personally I am happy that DARPA is making real headway in a field that has such a monumentious task. The perverbial road is miles long and we can only take baby steps, but the impacts of the work will be incalculable.
You can build the safest car in the world but there is always a need to be able to take a very quick decision to avoid some other idiot who might be breaking the rules of the road and not be in an automated car... still, if we all had them...
I am not sure if you are concerned that they will run the test on the open road (I doubt it) or about the prospect of robots on the open road in the future. Personally I think a degree of automatic control would even now prevent some of the really stupid behavior I see every day riding my bike to and from work.
Robot drivers will be somewhat better and somewhat worse than human drivers. This is true even today. It is nice that somebody is encouraging research.
What you had in mind was indeed the orriginal, often called Voltaire's dictum (sorry if I spelt that wrong) my sig is a somewhat nerdy joke about me considering that it is only my opinion which is worth dying for (for anyone) and that so long as I have freedom of speach it's all good... I suppose if I were to consider the nature of power I might find deeper ramifications to others not being able to speak out, I mean, if hegamonic marxism really is right then I might not notice one manifestation of the ideol
...how this can be "safe" (as far as it can be anyway) with cars which are automated going on roads for which the system works.
You can build the safest car in the world but there is always a need to be able to take a very quick decision to avoid some other idiot who might be breaking the rules of the road and not be in an automated car... still, if we all had them...
How is it any less safe? If anything, the feature with which human beings far outsurpress computers is our ability to filter huge amounts
You can build the safest car in the world but there is always a need to be able to take a very quick decision to avoid some other idiot who might be breaking the rules of the road and not be in an automated car... still, if we all had them...
I would argue that most people on the road are not able to do this while they drive. Whether they are actually capable if they are concentrating or anticipating is another matter.
If it were easy with would neither be a challenge, nor be grand.
You can build the safest car in the world but there is always a need to be able to take a very quick decision to avoid some other idiot who might be breaking the rules of the road and not be in an automated car... still, if we all had them...
If you read the RTFPR, there's a little blurb at the bottom that explains:
ABOUT THE DARPA GRAND CHALLENGE DARPA has sponsored two previous Grand Challenge competitions. The first was held in March 2004 and featured a 142-mile desert course. Fifteen autonomous ground vehicles attempted the course and no vehicle finished. In the 2005 Grand Challenge, four autonomous vehicles successfully completed a 132-mile desert route under the required 10-hour limit, and DARPA awarded a $2 million prize to "Stan
You can build the safest car in the world but there is always a need to be able to take a very quick decision to avoid some other idiot who might be breaking the rules of the road and not be in an automated car... still, if we all had them...
This is the Grand Challenge I was really waiting for. I believe that the experience gained in the previous Grand Challenges is practically useless for this new one. This new challenge will involve true AI, that is, AI that has true general learning capabilities and the
I'd be willing to bet these streets will be filled with drones or atleast stunt drivers will rollcages. They arn't going to run this experiment on an open street. Atleast I hope not!!
Either way I wonder if these teams will be allowed to use the existing lane following/adaptive cruise control technology that exist in some of the highend cars (yes I know lane following doesn't exist in the US right now, but the technology does exist)
Ahh, memories of my drive into work this morning and terrible human drivers [slashdot.org]. At the very least, robotic cars might be deterministic (depending on algorithms), right?
Because if your follow the basic rules of the road (and have cameras at all blind spots) these things are easy. What concerns me is that this being a challenge teams won't work together which means that the cars will not have a universal set of rules for communicating, something really essential in autonomous cars. (IE a radio network in which cars can say, I'm turning left, or I'm about to change into the left lane) Sure these cars will have blinkers as well, but that hugely ineffective compared to the pre
I'd rather they do it the hard way. Sure, it would make for an easier contest to have the cars communicate with each other; but if the goal here is to one day actually use this technology in the real world, how many kids on bicycles will have communication radios built into their bikes?
Good point. I was thinking that these cars should have radio communication AND the ability to use/recognize blinkers, etc. But if you assume that all communication you will receive will be minimal (ie drive like everyone might do something unexpected) your overall success in dealing with problems will be greater. Still would love to see an communication system between cars.. Many are the days I've seen a wreck about to happen that the drivers couldn't see and WISHED there was some effective way I could have
Its too early to go urban. They should have spent at least another 2-3 years perfecting autonomous navigation in unstructured environments.
I know last year's challange seemed to be won rediculously easily, but I have seen no proof that that dormain has been fully conquered yet. If they wanted a challenge why not move onto wooded or swampy areas.
In this case it seems they are juat setting themselves up to fail.
In this case it seems they are juat setting themselves up to fail.
No, they're setting a difficult task. Now look up the definition of "challenge". See?
People said the same thing after the first challenge, but people got the hang of that pretty quickly, and from what I've read over the last couple of years, this challenge should be just about doable.
Two reasons: First, because this is what the Army needs, right now. Imagine a supply convoy moving around Iraq without drivers, just combat guards. Fewer drivers=fewer IED deaths + soldiers who can be actual soldiers and MPs, not truck drivers. (Then again, having been a support platoon leader, I'm not sure I want some of those drivers holding a gun...)
Second, because it's a challenge! Last year's competition was not easy- look at the one two years ago, where most of the contestants barely made it out of the starting gate. (Or didn't) When proposed, it was an absurd reach- no robot had come anywhere close to the Grand Challenge specs; they were all busy managing 5-10 mph on easy courses. Nobody would have been interested if the challenge had been 10mph for 10 miles. Would they get the same number of entrants this year if the challenge was basically the same as before?
I agree, this is a serious reach. But honestly, it's not impossible. You have a *lot* more to navigate by in the city- all sorts of yellow and white paint lines on the road, existing high-detail maps and standardized road signs. Most of the drive will be free of serious obstacles, although I assume DARPA will throw in a some shell holes and road blocks to make it interesting. The radars to track other cars already exist- you can buy a car today with sophisticated cruise control that maintains correct distances. My guess is that robots will be better in city driving than humans very soon- 360 degree radar, much faster reaction times, no Starbucks latte and cell phone...
I suspect the next Grand Challenge will be something like "Start at Depot A, navigate across 300 miles of varying terrain, drop off at city center B". Speaking as that ex-platoon guy, most of the drivers in my unit couldn't *read*, much less read a map. They had to be led the entire way by someone who could. (Or at least, pretended he could- this was in the days before GPS.) Bring on the robots.
That's exactly right. I ran one of the Grand Challenge teams, Team Overbot [overbot.com], and we made it to the NQE. It's clear where DARPA is going, and they're getting there faster than they expected. There were 43 autonomous vehicles at the NQE, and all of them more or less worked. Five finished the course, and most of the 23 that started the course probably could have finished with minor improvements. This is way ahead of anything previously seen in robotics.
The big challenge this time is that now real situati
Just 1 team has to "win", and win doesn't have to be a complete success. If they can demonstrate the technology is 70, 80, even 90% there then that technology can be taken from the contestants and worked on in industry.
Let's make this real, vehicles must be equipped to :
Avoid hub-cap thieves.
Deal with homeless window-washers.
Handle crazy taxi-cab drivers.
Deal with traffic detours and malfunctioning signals.
Understand parking, loading, no stopping and no standing zones.
Not run over pedestrians who jaywalk.
Avoid accident scences/traffic congestion by planning on-the-fly alternate routes.
Be able to pay Tolls.
Stop, Look, Listen at Railroad crossings.
Be able to do this in even in Snow conditions.
Properly prepared, an urban course will be MORE challenging than the Desert course. Sure terrian won't be such an issue but the environment is much more dynamic, and the weather is still a factor.
"Grand Challenge 2005 proved that autonomous ground vehicles can travel significant
distances and reach their destination, just as you or I would drive from one city to the next," said
DARPA Director Dr. Tony Tether. "After the success of this event, we believe the robotics
community is ready to tackle vehicle operation inside city limits."
I'm not going to speculate as to whether the robotics community is "ready" for this challenge, but what do the two challenges have to do with each other from a technical standpoint? In the previous challenges, vision wasn't good enough to tell a boulder from a bush. Are they going to give the robots the GPS location of all the stop signs and traffic circles? If they do, how well would this apply to a city where not all GPS locations are known? If not, how will it differentiate signs from one another and from random stuff in the background?
I'll be impressed with no crashing into each other, before they worry about compliance with all traffic laws. How will the robots recognize the speed limit in their area, or will they all crawl along at 10 mph, impeding the flow of traffic?
I'll be impressed with no crashing into each other, before they worry about compliance with all traffic laws. How will the robots recognize the speed limit in their area, or will they all crawl along at 10 mph, impeding the flow of traffic?
Um, that's why they call it a "challenge" - because they don't know how to solve the problems yet. If it weren't really hard there would be no need to do this sort of contest.
Are they going to give the robots the GPS location of all the stop signs and traffic circles? If they do, how well would this apply to a city where not all GPS locations are known? If not, how will it differentiate signs from one another and from random stuff in the background?
The point of these challenges isn't to set one-year goals. An urban enviornment sets up a hugely more complicated affair that will requires years of failure before success. The complexity of the task goes up an order of magnitude.. however you are definitely hung up on the wrong problems. Signs occur at predictable locations, move in predictable ways, have predictable shapes, and use predictable colors. Someone with an introductory graduate course in computer vision could write a "sign" detector that is pretty robust.
but what do the two challenges have to do with each other from a technical standpoint?
When you boil it down, they're the exact same thing -- this is just a couple orders of magnitude more difficult. The previous challenge didn't have them dealing with any dynamic variables -- no passing vehicles, no being passed by vehicles, no boulders rolling off a mountain, etc.
And if you're going to solve those problems, why not do it for real? A boulder falling off the side of the road is reasonably uncommon. A car cutting you off is not (n.b. -- the challenge doesn't actually talk about this as an issue, and it may not be; we'll know more after May 20).
It's still all about road detection, object detection, and avoidance. And you're asking what they have to do with each other technically?
Are they going to give the robots the GPS location of all the stop signs and traffic circles?
Again, we won't know until after the Participant Conference on May 20, but I'd actually suspect they will, along with info on what speed limits apply in different areas (as they did last time). This is not unreasonable -- GPS mapping a city is pretty trivial when it comes down to it, and I doubt that the challenge is geared toward being fully dynamic -- e.g. you'll still follow a predetermined route, there won't be sudden changes in traffic rules (no road crews), and so forth.
That said, even if you have full GPS info on stop signs and so forth the most that's useful for is that you need to be watching out for a sign coming up soon. GPS isn't accurate enough (at least on a moving vehicle) to rely on it for road signs -- coming to a complete stop 3m beyond the stop sign doesn't work so well. So they'll still have to visually recognize a lot of traffic signage.
In some ways this will be easier than the previous challenge -- this is all low speed, so the issue of not being able to process the incoming data in real time will be reduced. On the flip side, you'll have to process a lot more data this time -- as you said, you must be able to recognize the difference between a boulder and a bush for this challenge.
I'll be impressed with no crashing into each other, before they worry about compliance with all traffic laws.
I'll be absolutely stunned if anyone succeeds this year, and moderately surprised if anyone succeeds at the one after.
But once this is complete, on to the next challenge -- mixed mode driving (urban, suburban, highway, maybe offroad). Then you can't tailor your algorithm toward a specific goal.
I wonder if you get extra points for having your car honk its horn at other drivers and extend it's little robotic middle finger at anyone that gets in its way?
... if they have alert drivers in the other cars, and... if they select an urban area which is in decent repair, and... if they allow ideal driving conditions like the last race. Think about it. City driving is designed to be easy. In fact it is really really easy. You are told exactly where to go with visible lines, lights, signs, etc which are all designed to be noticed and easily intepreted.
The hardest part of GC1 was finding the road! When it's layed out for you nice and easy.... man thats a cakewalk.
Why is it that every time I hear a story about efforts to improve vehicles, I say to myself, "Gee, hasn't that been done already?" All these efforts seem to have one thing in mind: get a car to act like a train, that way we can continue subsidising the auto/oil/rubber industries with the needless purchase of more individual rail cars.
How about a challenge to develop real public transit in the U.S.?
Oh, right, because you are a troll and couldn't give a rat's ass about the plight of others. This is the oldest troll in the book: some cool new thing comes along and some asshat who has never helped another human being in his life says, "What a waste of time! What about all the starving orphans in Africa?" or some such shit. Go crawl back under your bridge, troll, and let the grown ups talk.
Is that "help people" in the libertarian sense of "any goddam selfish thing I do helps people," because selfishness and greed are good? Or do you really help people? (in which case, my apologies, and carry on with your criticisms, you've earned the right)
My Dissapointment in DARPA (Score:5, Insightful)
Yes, they do research in defense but shouldn't there be a little more than a tiny graphic or blurb about what work they're doing? Couldn't they at least take the time to write an abstract or 1-2 page paper with unclassified information on each project?
Instead, I find the following links in the 'Archives' [darpa.mil]:
- Quantum Computing [darpa.mil]
- Infrared Focal Plane Array/Uncooled Integrated Sensors [darpa.mil]
- Advanced Lithography [darpa.mil]
- Most of the Other Links [darpa.mil]
My alma mater has produced better papers than this in these fields. I know that a lot of this stuff isn't classified and they list their programs on their sites, why can't they do a better job in showing the American public what they've done with our money?The Grand Challenge Forums [darpa.mil] are flooded with only vendors. Where are the designs and reports by the teams from older Grand Challenges? Why isn't this structured more like RoboCup where the learning algorithms are released every year so that future contestants can build on this?
The fact that this contest isn't run in a more open way makes it seems like less of a "contest" and more of a "do our research for us!" kind of thing.
Re:My Dissapointment in DARPA (Score:2)
forgive me but I think that is the whole point of these kind of contests.
It is likely that DARPA has become incapable of inovation because of internal politics so they need to attract new ideas from the out side.
This is all just my theory so take it with a grain of salt.
My Dissapointment in accuracy. (Score:2, Informative)
Um, Darpa is a think tank. They don't do actual research. It's been that way since the beginning.
Re:My Dissapointment in DARPA (Score:5, Informative)
Parent
Re:My Dissapointment in DARPA (Score:3, Insightful)
You'd rather find out DARPA has handed out a $10B contract to a "regular" defense contractor to solve this particular problem in ten years, only to find out at year #9, they will need a seven year extension and have cost overruns of another $8B?
A couple of years, a few million dollars as one carrot, the other two are establishing legitimacy in a captitalistic market looking for established technology, and finally, chest thumping in the geek world is a very, very tiny investment by comparison.
Wouldn't
Re:My Dissapointment in DARPA (Score:5, Informative)
When you say your alma mater "has produced better papers in these fields" you should have a look at the acknowledgements section of these papers. Chances are pretty good many of them will have a statement like "This work funded in part by DARPA (or NSF, etc) grant number XXX."
Parent
do our research for us (Score:2)
Re:My Dissapointment in DARPA (Score:3, Informative)
Well, yes.
This ISN'T about open technologies, this is about building up to a defense (warmaking) capability.
On the one hand, you have to release a tiny bit of information, just so all the competitors have the same basic assumptions and stay in the same universe of solutions. On the other hand, you don't want to constrain the competitors too much with regard
Whilst I understand.... (Score:4, Insightful)
We believe the robotics community is ready to tackle vehicle operation inside city limits. - Dr. Tony Tether, DARPA Director
You can build the safest car in the world but there is always a need to be able to take a very quick decision to avoid some other idiot who might be breaking the rules of the road and not be in an automated car... still, if we all had them...
Re:Whilst I understand.... (Score:4, Insightful)
You would needs some serious AI and pattern recognition to really replace drivers. There is just too much that can go wrong.
Parent
Re:Whilst I understand.... (Score:2)
Re:Whilst I understand.... (Score:2)
Re:Whilst I understand.... (Score:2)
Re:Whilst I understand.... (Score:2)
Re:Whilst I understand.... (Score:4, Insightful)
I am not sure if you are concerned that they will run the test on the open road (I doubt it) or about the prospect of robots on the open road in the future. Personally I think a degree of automatic control would even now prevent some of the really stupid behavior I see every day riding my bike to and from work.
Robot drivers will be somewhat better and somewhat worse than human drivers. This is true even today. It is nice that somebody is encouraging research.
Parent
Re:Whilst I understand.... (Score:2)
You disagree with everything I just said, but I will defend to the death my right to say it
in my mind should be:
You disagree with everything I just said, but I will defend to the death your right to say it.
Re:Whilst I understand.... (Score:2)
Re:Whilst I understand.... (Score:3, Insightful)
You can build the safest car in the world but there is always a need to be able to take a very quick decision to avoid some other idiot who might be breaking the rules of the road and not be in an automated car... still, if we all had them...
How is it any less safe? If anything, the feature with which human beings far outsurpress computers is our ability to filter huge amounts
Re:Whilst I understand.... (Score:2)
I would argue that most people on the road are not able to do this while they drive. Whether they are actually capable if they are concentrating or anticipating is another matter.
If it were easy with would neither be a challenge, nor be grand.
Oh and don't worry human,
Re:Whilst I understand.... (Score:2)
That's where the D in DARPA comes in.
Re:Whilst I understand.... (Score:2, Informative)
If you read the RTFPR, there's a little blurb at the bottom that explains:
ABOUT THE DARPA GRAND CHALLENGE DARPA has sponsored two previous Grand Challenge competitions. The first was held in March 2004 and featured a 142-mile desert course. Fifteen autonomous ground vehicles attempted the course and no vehicle finished. In the 2005 Grand Challenge, four autonomous vehicles successfully completed a 132-mile desert route under the required 10-hour limit, and DARPA awarded a $2 million prize to "Stan
True AI (Score:3, Insightful)
This is the Grand Challenge I was really waiting for. I believe that the experience gained in the previous Grand Challenges is practically useless for this new one. This new challenge will involve true AI, that is, AI that has true general learning capabilities and the
Re:Whilst I understand.... (Score:3, Funny)
Re:Whilst I understand.... (Score:2)
Either way I wonder if these teams will be allowed to use the existing lane following/adaptive cruise control technology that exist in some of the highend cars (yes I know lane following doesn't exist in the US right now, but the technology does exist)
Hardly fair... (Score:5, Funny)
Re:Hardly fair... (Score:2)
Re:Hardly fair... (Score:3, Interesting)
(IE a radio network in which cars can say, I'm turning left, or I'm about to change into the left lane) Sure these cars will have blinkers as well, but that hugely ineffective compared to the pre
Re:Hardly fair... (Score:3, Informative)
Re:Hardly fair... (Score:2)
Too early to go Urban. (Score:2, Insightful)
Its too early to go urban. They should have spent at least another 2-3 years perfecting autonomous navigation in unstructured environments.
I know last year's challange seemed to be won rediculously easily, but I have seen no proof that that dormain has been fully conquered yet. If they wanted a challenge why not move onto wooded or swampy areas.
In this case it seems they are juat setting themselves up to fail.
Surur
Re:Too early to go Urban. (Score:3, Insightful)
In this case it seems they are juat setting themselves up to fail.
No, they're setting a difficult task. Now look up the definition of "challenge". See?
People said the same thing after the first challenge, but people got the hang of that pretty quickly, and from what I've read over the last couple of years, this challenge should be just about doable.
Re:Too early to go Urban. (Score:4, Insightful)
Second, because it's a challenge! Last year's competition was not easy- look at the one two years ago, where most of the contestants barely made it out of the starting gate. (Or didn't) When proposed, it was an absurd reach- no robot had come anywhere close to the Grand Challenge specs; they were all busy managing 5-10 mph on easy courses. Nobody would have been interested if the challenge had been 10mph for 10 miles. Would they get the same number of entrants this year if the challenge was basically the same as before?
I agree, this is a serious reach. But honestly, it's not impossible. You have a *lot* more to navigate by in the city- all sorts of yellow and white paint lines on the road, existing high-detail maps and standardized road signs. Most of the drive will be free of serious obstacles, although I assume DARPA will throw in a some shell holes and road blocks to make it interesting. The radars to track other cars already exist- you can buy a car today with sophisticated cruise control that maintains correct distances. My guess is that robots will be better in city driving than humans very soon- 360 degree radar, much faster reaction times, no Starbucks latte and cell phone...
I suspect the next Grand Challenge will be something like "Start at Depot A, navigate across 300 miles of varying terrain, drop off at city center B". Speaking as that ex-platoon guy, most of the drivers in my unit couldn't *read*, much less read a map. They had to be led the entire way by someone who could. (Or at least, pretended he could- this was in the days before GPS.) Bring on the robots.
Parent
Re:Too early to go Urban. (Score:3, Informative)
The big challenge this time is that now real situati
Not everyone has to win (Score:2)
Do they have to... (Score:4, Funny)
Re:Do they have to... (Score:2)
readiness? (Score:5, Interesting)
I'm not going to speculate as to whether the robotics community is "ready" for this challenge, but what do the two challenges have to do with each other from a technical standpoint? In the previous challenges, vision wasn't good enough to tell a boulder from a bush. Are they going to give the robots the GPS location of all the stop signs and traffic circles? If they do, how well would this apply to a city where not all GPS locations are known? If not, how will it differentiate signs from one another and from random stuff in the background?
I'll be impressed with no crashing into each other, before they worry about compliance with all traffic laws. How will the robots recognize the speed limit in their area, or will they all crawl along at 10 mph, impeding the flow of traffic?
Re:readiness? (Score:2, Insightful)
Re:readiness? (Score:4, Informative)
The point of these challenges isn't to set one-year goals. An urban enviornment sets up a hugely more complicated affair that will requires years of failure before success. The complexity of the task goes up an order of magnitude.. however you are definitely hung up on the wrong problems. Signs occur at predictable locations, move in predictable ways, have predictable shapes, and use predictable colors. Someone with an introductory graduate course in computer vision could write a "sign" detector that is pretty robust.
Parent
Re:readiness? (Score:4, Insightful)
When you boil it down, they're the exact same thing -- this is just a couple orders of magnitude more difficult. The previous challenge didn't have them dealing with any dynamic variables -- no passing vehicles, no being passed by vehicles, no boulders rolling off a mountain, etc.
And if you're going to solve those problems, why not do it for real? A boulder falling off the side of the road is reasonably uncommon. A car cutting you off is not (n.b. -- the challenge doesn't actually talk about this as an issue, and it may not be; we'll know more after May 20).
It's still all about road detection, object detection, and avoidance. And you're asking what they have to do with each other technically?
Are they going to give the robots the GPS location of all the stop signs and traffic circles?
Again, we won't know until after the Participant Conference on May 20, but I'd actually suspect they will, along with info on what speed limits apply in different areas (as they did last time). This is not unreasonable -- GPS mapping a city is pretty trivial when it comes down to it, and I doubt that the challenge is geared toward being fully dynamic -- e.g. you'll still follow a predetermined route, there won't be sudden changes in traffic rules (no road crews), and so forth.
That said, even if you have full GPS info on stop signs and so forth the most that's useful for is that you need to be watching out for a sign coming up soon. GPS isn't accurate enough (at least on a moving vehicle) to rely on it for road signs -- coming to a complete stop 3m beyond the stop sign doesn't work so well. So they'll still have to visually recognize a lot of traffic signage.
In some ways this will be easier than the previous challenge -- this is all low speed, so the issue of not being able to process the incoming data in real time will be reduced. On the flip side, you'll have to process a lot more data this time -- as you said, you must be able to recognize the difference between a boulder and a bush for this challenge.
I'll be impressed with no crashing into each other, before they worry about compliance with all traffic laws.
I'll be absolutely stunned if anyone succeeds this year, and moderately surprised if anyone succeeds at the one after.
But once this is complete, on to the next challenge -- mixed mode driving (urban, suburban, highway, maybe offroad). Then you can't tailor your algorithm toward a specific goal.
Parent
Detroit area (Score:2)
I wonder if... (Score:4, Funny)
I wonder if you get extra points for having your car honk its horn at
other drivers and extend it's little robotic middle finger at anyone that
gets in its way?
The real reason they're doing this. (Score:3, Interesting)
Re:The real reason they're doing this. (Score:2)
Better still: Unmanned suicide bombers!
This will be easy ... (with caveats) (Score:2)
Think about it. City driving is designed to be easy. In fact it is really really easy. You are told exactly where to go with visible lines, lights, signs, etc which are all designed to be noticed and easily intepreted.
The hardest part of GC1 was finding the road! When it's layed out for you nice and easy.... man thats a cakewalk.
D
Public Transit, anyone? (Score:2, Interesting)
Why is it that every time I hear a story about efforts to improve vehicles, I say to myself, "Gee, hasn't that been done already?" All these efforts seem to have one thing in mind: get a car to act like a train, that way we can continue subsidising the auto/oil/rubber industries with the needless purchase of more individual rail cars.
How about a challenge to develop real public transit in the U.S.?
Sorry, I live in L.A. and I'm bitter.
Why not quit your job and feed the homeless? (Score:2)
Re:Why not quit your job and feed the homeless? (Score:2)
Re:Why not quit your job and feed the homeless? (Score:2)
As a co-worker of mine likes to say: (Score:2)