"To the best of my knowledge, it has not been published how they learned the far range vision based on the near range laser scanner, which, to my eyes, is the most interesting part of the project."
They didn't use the near range laser scanner for the far range vision. They had a color camera they used. That was what was used to get those Red/Green video's/pictures you see on Nova's site. None of the competitors used the SICK Lasers for anything more than 25 meters, because they simply didn't work past that. A couple of teams had a different laser, but I can't remember the name, I know it was around $30-35k a pop, vs the $7500 a pop for the SICK. A little far out of our price range. Stanford also had RADAR, but AFAIK that was only for range calculations and suplimental drivability map data.
Camera whether stereo or not was the big "if" in the competition. It seemed like it was a surefire way to doom your team. We tried video, but in the end we ended up relying on LADAR like everyone else. To my knowledge even Standford only slighly used their camera system, compared to their LADAR/RADAR use. It was just too hard to get working reliably. Given another year, we could have had ours trained to drive you to work, but like everyone else, there simply wasnt the time available to train a vision system.
Nice try, I wasn't on CMU either. No I was on a team that actually had to work for every dime they spent on the vehicle.
I'm not stating anything false. Read his research and when it took place. The core stuff all came from CMU or Sebastian in previous efforts (like I stated). Now once at Stanford they changed how they did things entirely and wrote a ton of code to make everything play much nicer than CMU's platform. Stanly was much smarter than either of the two CMU bots.
Honestly, not really. It was so damn dry out there that they water would spray the dust off and dry off in no time. I'd say rarely though did we ever see the water system turn on. Really, only in our mud testing did we ever get major buildup. Those LADAR's were pretty resilient sensors. The sun shining in them was much worse than any dust buildup.
how hard would it be for a motivated, talented individual to figure out some simple attacks for the navigation systems on these vehicles, and get loads of sweet US munitions delivered to their doorstep?
How hard? I would say next to impossible. Tricking the GPS system on board is impossible, while staying alive. You'd have to emulate enough GPS sats to give the vehicle improper coordinates (not only would these Satellite emulators have to work, but they would have to be synchronized properly, don't forget to emulate Doppler effects!). Even if you do trick the vehicle, you'll have to jam the other GPS satellites to make sure you are the signal it goes for. If you try jamming GPS, you'll get a HARM sent your way (bad news). If you are talking attacking the ACTUAL navigation system on board, that means you have physical access to the vehicle, which means there's no point for it to navigate to your house, you can just offload the goods. In that case, I bet they'd either let you have the goods (if it's water or something) or drop a bomb on you while you are unloading it (if it's something more sensitive).
Your next two questions are much more valid. Urban setting will be tough. The only benefit is at least there will be some form of road boundry to work off of. The open desert is much tougher in that regard, everything looks the damn same. I assume these systems will begin to be used in the more long haul type scenarios. Open highway with no buildings, but are really dangerous to drive because of ambushes and IED's.
Many of the teams (including us, small team, but still finalist) had little water spray nozzles. The LADAR would actually have a very specific error code when it was being disrupted, so we triggered on that. We also disregarded all "obstacles" within 1 cm of the sensor (ie: the lense!).
Although software was a big focus, having good mechanical engineers was the key. Keeping stuff cool and free from vibration type failures was a BIG deal.
Do a google search on Sabastian Thrun, he was the team lead for Stanford, and formally at CMU (what a non-coincidence). Most of the software they used on Stanly (Stanford's bot) was either written by Sebastian in his former research or taken from experience gained on CMU's team the previous year. The ladar mapping he used, I know I saw on some former page of his that had all the gory algorithm details. It might just take a little bit of searching. He also has a c library out there somewhere that does a lot of this stuff, but I can't seem to find it now.
"He also pointed to statistics that show the United States graduating only 4,400 mathematics and science PhDs each year compared with 24,900 math and science PhDs for greater Asia."
So let me get this straight. The U.S.A. has a population of 275 million or so and graduates 4,400 math and science PhDs, while greater Asia (just what is included in that, I have no idea??) with a population of 1 or 2 billion and graduates 24,900 math and science PhDs.
Where's the problem? It seems like we are doing fine in that regard.
Sorry, I forgot to bookend my post with tags. I can't believe it. I was selfishly mentioning the saving of U.S. lives and only because we are talking about a U.S. Defense competition. Jesus, what was I thinking.
<\america_is_evil>
Next time you reply, can you at least bookend your post with <vagina> tags?
I don't know about these being used as unmanned 'weapons', but certainly the military will use them for transport vehicles. The real reason for this competition was to create technology that would save lives. This is even more appropriate now that a majority of deaths over in Iraq are due to road side bombs. Right now, weapon control systems will still have a human somewhere in the loop.
No offense to them, but they lied about their run then. They drove for a solid 15ft on top of a pile of haybales when they ran into them and proceded to "stay the course" while on top of them. But maybe they left that off of their webpage (I didn't look, just taking your info). They still had a good run in my opinion, but they certainly did more than nudge obstacles. I'm not a fan of any of the big teams, but Stanfords run really has been the best so far.
I realize that. My point was, one of the contributing factors for the push for UAV's was a replacement for the aging U2. Not only were its capabilities fading, you had to rely on a human to endure horrible lengths of time up in the air over hostile territory. The UAV allowed that requirement to disappear. The government has a huge Christmas list of the stuff they want/need. They only have enough money to do a small subset of them. My only point was that the government needs an effective spy vehicle, the UAV provided that. My post was to be taken in jest, obviously. No one would think that the Humvee would be a good spy vehicle, the point was to say that a good spy vehicle was higher on the Christmas list than a unmanned ground vehicle to prevent casualities. The OP was most right in saying that UAV's are somewhat more simple because they don't have objects to run into. I do argue though that ground vehicles can *stop* when they are confused about something, a UAV doesn't have that luxury. The best it can do is circle over a location until it becomes reaquainted with its mission.
Yes, but it's also because the government REALLY wanted Autonomous air vehicles. Humvees don't make the best spy vehicle. A plane that can fly for 36 hours over a location taking pictures on the other hand is *very* handy.
Yeah buddy, it's *that* hard. If you don't think so, please tell me how (using a laser or camera) to have a computer tell you the difference between a canyon (no laser return), a water puddle (no laser return, reflection on the camera), and the sky (ie: going through whoops and the vehicle is looking straight up). Yes, a gyro can help you with the latter. Our mind can figure out a lot of things that just take too much time to do with computers. I can see a puddle and look to either side of it to see how the terrain dips, then analyzing the width and length of the puddle, I can guess if it is shallow enough to be crossed. All of this can be thought of in 4 or 5 seconds in my mind (just fast enough to make the decision before hitting it at 20mph), but a supercomputer doesn't stand a chance.
Plus, before mentioning anything about difficulty, try driving offroad first. Our team would have been done 8 months ago if we were sticking to paved roads. Personally I drive all the time in Baja, and going any faster than about 35 or 40 mph scares the crap out of me.
I'm sure there will be more than a few teams that do better than the 7 miles last year. I'm on one of the teams and we have done many miles fully autonomous in the Anza Borrego Desert (very similar to the conditions at the DGC). The NQE is going well, on the first day many teams passed on their opportunity for their first run becaues they weren't ready. Of the teams that did do their run, about half made it and half didn't. There were a couple of highlights, one of the favorite team's vehicle flattened a whole section of haybales, and at least one team participated in a full-on car accident with one of the parked car obstacles. A couple of teams did the course perfectly at slow speed.
I recommend that if any of you follow the race to pay particular attention to those teams who aren't getting in the papers. I'm most impressed by the teams who aren't university based. Most of them have little sponsorship and are just working guys spending their *own* cash on their vehicle. I purposely didn't name names of any teams because so many posts about the DGC have become spamvertisements here on Slashdot. *cough* CMU *cough* Oops.
They talk about the program right there. It took me 2-clicks to get there from your link. I agree they don't go into a lot of details, but they explicitly talk about reviving dogs.
Opera is available on the competitor's phone (Qualcomm) on the Brew platform. So, I don't suppose Nokia made it better, they just didn't have Opera available to them.
My team's (AVS) vehicle is a street legal (although mean looking Humvee) and got pulled over by police. They mostly wondered what the hell we were doing wondering around in a military Humvee on usually unused roads with tank-stopper obstacles in the bed. They were impressed with our explanation, wished us luck, and went on their way.
"I dont understand why their team is getting this much press"
I was thinking the same thing. My team has spent fractions (small fractions) of what these big University teams have spent, and we did really good during the site visit. Maybe next time, slashdot can post a more general DARPA GC post rather than some advertisement for a single team. *NEWSFLASH* CMU and Stanford don't need ANY additional sponsors.
As tempting it is to pimp my team, I won't, but just look around the GC site and read about some of the truly cool teams out there.
I guess I should have RTFA better the first time. Support for the "his wife is Chinese" is here:
"as someone who married into a Shanghai family, I congratulate the bright, dedicated members of the winning Jiaoda team, which also took first place in 2002"
I went to UCDavis, and all the students I knew loved Norm Matloff. He speaks Chinese, he was one of the first to do heavy research on supporting Chinese characters in software, and if I recall correctly, his wife is Chinese (I couldn't find it anywhere on his webpage to back that part up).
I hate to use the classic "but I have lots of black friends!" anti-racist argument here, but I think he has earned it. I think the reason your friends don't want him as their advisor is because he is one of the toughest Prof's at Davis, and he isn't going to give out a free ride through grad school.
Of course, you have been modded up, and no one is going to read my reply, so the false prejudiced accusation is what people will see. But again, this IS slashdot. The first to respond is always right!
As an aside, he was also a big reason that Intel Corp. in Sacramento changed their stance on G.P.A. being the major deciding factor in hiring a student. They used to throw out all resumes that were under a 4.0 G.P.A. (they had THAT many applying). Dr. Matloff basically showed them that the students that could REALLY program weren't the ones getting A's. He has a paper somewhere on his site, but again, no one is going to read this reply anyway!
Um, did you click on that link??? It was a link to a urban myth site that debunked the 10% rumor. I guess you use less of your brain than you advertised.
Slashdot Mirror
"To the best of my knowledge, it has not been published how they learned the far range vision based on the near range laser scanner, which, to my eyes, is the most interesting part of the project."
They didn't use the near range laser scanner for the far range vision. They had a color camera they used. That was what was used to get those Red/Green video's/pictures you see on Nova's site. None of the competitors used the SICK Lasers for anything more than 25 meters, because they simply didn't work past that. A couple of teams had a different laser, but I can't remember the name, I know it was around $30-35k a pop, vs the $7500 a pop for the SICK. A little far out of our price range. Stanford also had RADAR, but AFAIK that was only for range calculations and suplimental drivability map data.
Camera whether stereo or not was the big "if" in the competition. It seemed like it was a surefire way to doom your team. We tried video, but in the end we ended up relying on LADAR like everyone else. To my knowledge even Standford only slighly used their camera system, compared to their LADAR/RADAR use. It was just too hard to get working reliably. Given another year, we could have had ours trained to drive you to work, but like everyone else, there simply wasnt the time available to train a vision system.
Nice try, I wasn't on CMU either. No I was on a team that actually had to work for every dime they spent on the vehicle.
I'm not stating anything false. Read his research and when it took place. The core stuff all came from CMU or Sebastian in previous efforts (like I stated). Now once at Stanford they changed how they did things entirely and wrote a ton of code to make everything play much nicer than CMU's platform. Stanly was much smarter than either of the two CMU bots.
Honestly, not really. It was so damn dry out there that they water would spray the dust off and dry off in no time. I'd say rarely though did we ever see the water system turn on. Really, only in our mud testing did we ever get major buildup. Those LADAR's were pretty resilient sensors. The sun shining in them was much worse than any dust buildup.
How hard? I would say next to impossible. Tricking the GPS system on board is impossible, while staying alive. You'd have to emulate enough GPS sats to give the vehicle improper coordinates (not only would these Satellite emulators have to work, but they would have to be synchronized properly, don't forget to emulate Doppler effects!). Even if you do trick the vehicle, you'll have to jam the other GPS satellites to make sure you are the signal it goes for. If you try jamming GPS, you'll get a HARM sent your way (bad news). If you are talking attacking the ACTUAL navigation system on board, that means you have physical access to the vehicle, which means there's no point for it to navigate to your house, you can just offload the goods. In that case, I bet they'd either let you have the goods (if it's water or something) or drop a bomb on you while you are unloading it (if it's something more sensitive).
Your next two questions are much more valid. Urban setting will be tough. The only benefit is at least there will be some form of road boundry to work off of. The open desert is much tougher in that regard, everything looks the damn same. I assume these systems will begin to be used in the more long haul type scenarios. Open highway with no buildings, but are really dangerous to drive because of ambushes and IED's.
Many of the teams (including us, small team, but still finalist) had little water spray nozzles. The LADAR would actually have a very specific error code when it was being disrupted, so we triggered on that. We also disregarded all "obstacles" within 1 cm of the sensor (ie: the lense!).
Although software was a big focus, having good mechanical engineers was the key. Keeping stuff cool and free from vibration type failures was a BIG deal.
Do a google search on Sabastian Thrun, he was the team lead for Stanford, and formally at CMU (what a non-coincidence). Most of the software they used on Stanly (Stanford's bot) was either written by Sebastian in his former research or taken from experience gained on CMU's team the previous year. The ladar mapping he used, I know I saw on some former page of his that had all the gory algorithm details. It might just take a little bit of searching. He also has a c library out there somewhere that does a lot of this stuff, but I can't seem to find it now.
u blic_html/papers/thrun.ces-tr.html (sorry, no linky, writing in a hurry)
/ bfl-trunk/
h tml
One paper that's of interest might be here: http://www.cs.cmu.edu/afs/cs.cmu.edu/user/thrun/p
And that paper is mentioned in the readme of the BFL (Bayesian Filtering Library) found here:
http://people.mech.kuleuven.be/~kgadeyne/software
Lastly, at one point all of us competitors were required to give our design documents to DARPA, and they put them up on their webpage here:
http://www.darpa.mil/grandchallenge05/techpapers.
BTW, I wasn't on Stanford's team, but I was on another finalist team.
Did the beginning of the article:
"According to a pdf on Adobe's website..."
not give it away? I thought it was quite obvious. But I guess complaining is easier than reading.
"He also pointed to statistics that show the United States graduating only 4,400 mathematics and science PhDs each year compared with 24,900 math and science PhDs for greater Asia."
So let me get this straight. The U.S.A. has a population of 275 million or so and graduates 4,400 math and science PhDs, while greater Asia (just what is included in that, I have no idea??) with a population of 1 or 2 billion and graduates 24,900 math and science PhDs.
Where's the problem? It seems like we are doing fine in that regard.
Idiot.
I thought the same thing. When I saw the topic, I said to myself "SWEET!".
Then damn it all, everyone on slashdot in their IANAL garb squashed it.
Sorry, I forgot to bookend my post with tags. I can't believe it. I was selfishly mentioning the saving of U.S. lives and only because we are talking about a U.S. Defense competition. Jesus, what was I thinking.
<\america_is_evil>
Next time you reply, can you at least bookend your post with <vagina> tags?
I don't know about these being used as unmanned 'weapons', but certainly the military will use them for transport vehicles. The real reason for this competition was to create technology that would save lives. This is even more appropriate now that a majority of deaths over in Iraq are due to road side bombs. Right now, weapon control systems will still have a human somewhere in the loop.
No offense to them, but they lied about their run then. They drove for a solid 15ft on top of a pile of haybales when they ran into them and proceded to "stay the course" while on top of them. But maybe they left that off of their webpage (I didn't look, just taking your info). They still had a good run in my opinion, but they certainly did more than nudge obstacles. I'm not a fan of any of the big teams, but Stanfords run really has been the best so far.
I realize that. My point was, one of the contributing factors for the push for UAV's was a replacement for the aging U2. Not only were its capabilities fading, you had to rely on a human to endure horrible lengths of time up in the air over hostile territory. The UAV allowed that requirement to disappear. The government has a huge Christmas list of the stuff they want/need. They only have enough money to do a small subset of them. My only point was that the government needs an effective spy vehicle, the UAV provided that. My post was to be taken in jest, obviously. No one would think that the Humvee would be a good spy vehicle, the point was to say that a good spy vehicle was higher on the Christmas list than a unmanned ground vehicle to prevent casualities. The OP was most right in saying that UAV's are somewhat more simple because they don't have objects to run into. I do argue though that ground vehicles can *stop* when they are confused about something, a UAV doesn't have that luxury. The best it can do is circle over a location until it becomes reaquainted with its mission.
Yes, but it's also because the government REALLY wanted Autonomous air vehicles. Humvees don't make the best spy vehicle. A plane that can fly for 36 hours over a location taking pictures on the other hand is *very* handy.
Yeah buddy, it's *that* hard. If you don't think so, please tell me how (using a laser or camera) to have a computer tell you the difference between a canyon (no laser return), a water puddle (no laser return, reflection on the camera), and the sky (ie: going through whoops and the vehicle is looking straight up). Yes, a gyro can help you with the latter. Our mind can figure out a lot of things that just take too much time to do with computers. I can see a puddle and look to either side of it to see how the terrain dips, then analyzing the width and length of the puddle, I can guess if it is shallow enough to be crossed. All of this can be thought of in 4 or 5 seconds in my mind (just fast enough to make the decision before hitting it at 20mph), but a supercomputer doesn't stand a chance.
Plus, before mentioning anything about difficulty, try driving offroad first. Our team would have been done 8 months ago if we were sticking to paved roads. Personally I drive all the time in Baja, and going any faster than about 35 or 40 mph scares the crap out of me.
I'm sure there will be more than a few teams that do better than the 7 miles last year. I'm on one of the teams and we have done many miles fully autonomous in the Anza Borrego Desert (very similar to the conditions at the DGC). The NQE is going well, on the first day many teams passed on their opportunity for their first run becaues they weren't ready. Of the teams that did do their run, about half made it and half didn't. There were a couple of highlights, one of the favorite team's vehicle flattened a whole section of haybales, and at least one team participated in a full-on car accident with one of the parked car obstacles. A couple of teams did the course perfectly at slow speed.
I recommend that if any of you follow the race to pay particular attention to those teams who aren't getting in the papers. I'm most impressed by the teams who aren't university based. Most of them have little sponsorship and are just working guys spending their *own* cash on their vehicle. I purposely didn't name names of any teams because so many posts about the DGC have become spamvertisements here on Slashdot. *cough* CMU *cough* Oops.
Really?
h s_sa4.html
http://www.safar.pitt.edu/content/programs/safar/
They talk about the program right there. It took me 2-clicks to get there from your link. I agree they don't go into a lot of details, but they explicitly talk about reviving dogs.
a grain of salt?
"We need a fourth law of Robotics: Stop Fingering My Wife"
Classic.
Opera is available on the competitor's phone (Qualcomm) on the Brew platform. So, I don't suppose Nokia made it better, they just didn't have Opera available to them.
My team's (AVS) vehicle is a street legal (although mean looking Humvee) and got pulled over by police. They mostly wondered what the hell we were doing wondering around in a military Humvee on usually unused roads with tank-stopper obstacles in the bed. They were impressed with our explanation, wished us luck, and went on their way.
"I dont understand why their team is getting this much press"
I was thinking the same thing. My team has spent fractions (small fractions) of what these big University teams have spent, and we did really good during the site visit. Maybe next time, slashdot can post a more general DARPA GC post rather than some advertisement for a single team. *NEWSFLASH* CMU and Stanford don't need ANY additional sponsors.
As tempting it is to pimp my team, I won't, but just look around the GC site and read about some of the truly cool teams out there.
I guess I should have RTFA better the first time. Support for the "his wife is Chinese" is here:
"as someone who married into a Shanghai family, I congratulate the bright, dedicated members of the winning Jiaoda team, which also took first place in 2002"
I went to UCDavis, and all the students I knew loved Norm Matloff. He speaks Chinese, he was one of the first to do heavy research on supporting Chinese characters in software, and if I recall correctly, his wife is Chinese (I couldn't find it anywhere on his webpage to back that part up).
Here's his Chinese software page:
http://heather.cs.ucdavis.edu/chinese.html
I hate to use the classic "but I have lots of black friends!" anti-racist argument here, but I think he has earned it. I think the reason your friends don't want him as their advisor is because he is one of the toughest Prof's at Davis, and he isn't going to give out a free ride through grad school.
Of course, you have been modded up, and no one is going to read my reply, so the false prejudiced accusation is what people will see. But again, this IS slashdot. The first to respond is always right!
As an aside, he was also a big reason that Intel Corp. in Sacramento changed their stance on G.P.A. being the major deciding factor in hiring a student. They used to throw out all resumes that were under a 4.0 G.P.A. (they had THAT many applying). Dr. Matloff basically showed them that the students that could REALLY program weren't the ones getting A's. He has a paper somewhere on his site, but again, no one is going to read this reply anyway!
Um, did you click on that link??? It was a link to a urban myth site that debunked the 10% rumor. I guess you use less of your brain than you advertised.