Hey engineers! (nifty competition)

[Guest Nic-Rebel450CA]

Here is a nifty competition that I was just reading about, looks like it is failing this year, but should be run again next year.

 

http://www.globalsecurity.org/org/news/2004/040311-robot-race.htm

 

Offers a $1 million prize for the winner. I am planning on doing some research with my programming and GPS etc to try making a guidance system (hey, if Java is good enough for NASA ) Not sure how much else I could do, but it should be a fun project.

 

Just thought there are probably some engineers on here that might wanna check into this

[akeizm]

Haha I saw that on the news, and laughed my butt off. Everyone was talkin it up and none of them got further than 15 metres!

[strotter]

Not true - the Carnegie Mellon University entry travelled 7.4 miles before dying (went off course & snapped an axle). Check out http://story.news.yahoo.com/news?tmpl=story&u=/ap/20040315/ap_on_re_us/robot_race_1. The interesting thing isn't how quickly they died, it was how far some of them got. Remember, these guys were going to have to average about 20 mph to finish the course in time. When it went off the road, the Carnegie Mellon Humvee was doing about 40.

 

It may be funny this year, but the next time they do it (probably in 2006) they'll be better. And the time after that, and the time after that. Make no mistake, we'll see robotically operated automobiles in our lifetime. Heck, there are already a number of systems that can drive on freeways with no external guidance systems at all - I think MIT (correct me?) drove one from Boston to L.A. last year.

 

Also, cars are the last non-robotically controlled form of transportation. Trains, commercial aircraft, commercial boats and shipping, they're all automatic. Human pilots are in commercial aircraft mostly for the peace-of-mind of the passengers. The Air Force has a number of robotic (not tele-operated, robotic) drones, and is working on a robotic fighter capable of delivering ordinance and even engaging in dog-fights. There was a Popular Science article about that one - the prototype is capable of pulling 20 g's if necessary, carries AIM 9's (sidewinder missiles) and may be able to use a 20mm nose cannon in close-on combat.

 

There was a also a robotic device the British were working on a number of years ago called a "Turtle". It was basically a light tank with no gun but carrying a nuclear bomb. It was set up to trigger if molested. You'd broadcast a warning before you sent it on its' way. Shoot a cannon at it, it goes off. Try to break into it, it goes off. Interrupt the signal it's receiving from home base, it goes off. The idea was that you'd send it into the middle of a city where it would just park and wait. It was sort of the ultimate negotiating tool. You could disrupt an entire country without firing a shot, with four of five of them in major cities. Can you imagine the panic as people tried to get 20 miles away from one of these things? I don't know if they're still working on it, but that's one of those things that'll keep you up at night.

 

Also, the Navy is working on these little "crab" looking things that you'd dump into the ocean off a beach you're going to invade. They'd find mines and either mark them, or wait patiently until they got a signal and set them off, just before the Marines came in. They have to be robotic because radio signals don't go through water.

 

There's all kinds of interesting robotic things going on right now. Not all of it's *good*, for sure, but all of it's interesting. As the computer systems get smarter and more reliable we'll be seeing a lot more. Interesting times.

[QWKDTSN]

The important thing is how far some of them got? 7.4 miles out of 200+ isn't very good... and that's the BEST that any of them did!

[mobythevan]

dissapointing from an engineering standpoint.

[wheelman]

If you really think about it though, 7.4 miles for one of these systems is very good. Think about what they are navigating through, this isn't a freeway system with boundaries that are marked very obviously. These things are navigating across a desert area with little or no roads and have to find their own way. I think it's very good.

 

I've worked as a programmer doing embedded equipment control work for most of the 20 years I've programmed and I can truthfully say the software on these things is very difficult to develop. Most of my embedded work has been with scientific lab equipment that lives in a controlled environment and the software gets very sophisticated. These units are dealing with a non-static environment through which they are asked to navigate at speed. That takes mucho processing power and a tough vehicle.

 

It's intriguing to speculate about where things will be in 20 or 30 years. Look at how much more processing power computers have today vs 10 or even 5 years ago (Moores law). Now extrapolate that into the future and we'll have computers monitoring and running everything for us, that is if we don't run out of energy before then. Well not in 20 or 30 years but go out 50 to 100 and new energy sources will have to be developed.

 

Wheelman

[TimZ]

Interestingly, my boss was one of the judges for this event. I was just talking with him about it last night.

 

This was a monumental task, and it was not at all surprising that nobody made it that far. The level of sophistication required to make a vehicle that could navigate it's own way from Las Vegas to Los Angeles is orders of magnitude higher than any automated mass transit system currently in use.

 

The contestants were not given the route until 2 hours before the race, and it consisted of about 2500 GPS waypoint coordinates that they had pass within some distance of to stay in the race. The vehicles are required to figure out how to get around whatever obstacles they might encounter on their own. The only external control allowed was for a kill switch so that is could be disabled if something went awry.

 

While it doesn't surprise me that people with no exposure to this technology might underestimate the magnitude of the problem at hand, I did find it pretty amazing that several of the teams chose to make custom vehicles, instead of adding hardware to an already proven propulsion package.

 

One team chose to add a special propane powertrain to their vehicle(I think it was a sand rail of some sort), and at the last minute figured out that they would not have enough range to make it on one tank. So they tore out the passenger seat and added a second propane tank. Apparently, this changed the suspension loading enough that their steering linkage started binding, and they couldn't even start the race. I just don't get it - why would you choose to add completely unneeded complexity to an already nearly impossible task?

 

Then there was the guy that decided it would be cool to use an autonomously guided motorcycle. He spent pretty much all of his time f-ing around trying to get the thing to not fall over after 10 or 15 feet. It sounded like he didn't even start on the navigation stuff.

[mobythevan]

I am still dissapointed. Very dissapointed over the hype and the results. Makes engineers look like a bunch of dummies(which may be true). I design and debug scientific computers and we have several hardware accelerators for image detection. I would expect more, but not for a million dollar prize. I have plenty of insight into this technology, but I can't imagine making it work(finish the race) for under a million. Oh well, another bad showing for engineers and technology. We'll see what happens next year.