Grand Challenge Results
Posted 14 Mar 2004 at 21:52 UTC by steve
Everyone has heard by now that none of the robots completed DARPA's
Grand Challenge yesterday. With the backing of major corporations like
Intel and Boeing, more than 3 million dollars and expert assistance from
the CMU Robotics Lab, the Red Team's robot managed 7 miles of the 142
mile course before
getting stuck on a rock and spinning the tires until the rubber caught
fire. Much smaller teams with less funding like the Golum Group and Team
DAD did just as well. Stories with details of who crashed and how can be
found in the Mercury
Vegas Review Journal, and plenty
of others. Photos can be found on Yahoo!,
DARPA Media Gallery
and elsewhere. And
the Register has a good story on
Team Phantasm's experiences.
First off, if it where such an easy task then DARPA wouldnt have held
the competition with such a large reward for the winner; maybe next
year they will allow the rest of us to participate rather than turning
away 3/4 of the entrants several months before the competition.
Participation, posted 15 Mar 2004 at 02:38 UTC by steve »
Yes, what I'd like to see more than details about how the 15 robots died
is a more detailed explanation of why each of the other teams were
disallowed from participation at all. Just getting an entry to the
starting line is a Grand Challenge in itself!
Participation, posted 15 Mar 2004 at 18:51 UTC by while_true »
The failure of so many teams to even get past the first 200 yards
proved that DARPA was right not to include many contestants that
didn't have an ice-cube's chance in hell of making it.
Just because they didn't disclose their rubric, doesn't mean their
pruning of contestants was random.
I personally think the next step should be to give a generous
operating budget (say $10 million) to each team that made it close to
a mile. The point is that the capabilities of hobbyist with day jobs
are limited. Let them quit their day jobs.
For $13 million in costs for this event, they got 7 miles from 2
teams. For $70 million, they might have the problem solved. Note that
this is still FAR cheaper than an average budget for a new vehicle, in
contracts to large defense firms.
Say what?, posted 16 Mar 2004 at 00:31 UTC by WhoPhlungPoo »
The spirit of the competition was originally intended to give everyone
interested in participation a chance; ABSOLUTELY NO ONE should have
been excluded from this event; regardless of weather or not you or
anyone within the DARPA organization felt they didn't have a chance;
they should have at bare minimum been allowed to go to the qualifying
event. Instead of changing the qualifying rules at the last minute so
they could actually hold the race with more than two participants; they
could have stuck with the original plan and let the qualifying event
weed out the entrants that didnt muster up; there are a lot of
creative people interested in doing something like this, just because
they aren't associated with a prestigious school like CMU or similar
group doesnt mean they don't have a chance. Remember those two bicycle
mechanics that kicked off the entire powered flight industry?
Creativity can not be learned, purchased or validated by a college
degree; you're either born with it or you're
say this..., posted 16 Mar 2004 at 01:35 UTC by while_true »
Interesting reply, but keep in mind:
1) The rules weren't changed during the QID to allow the other
contenders in. There was never a stipulation that passing the QID was
required to enter the race. Correct me if I am wrong here.
2) There is nothing stopping a anyone from going to the desert right
now and repeating the race along the same way-points. If you think you
can do it, go ahead and try. If you succeed, there will be at least a
dozen groups interested in buying your technology, which is worth well
over $1M, and everyone knows it. Hell, I'll give you a million dollars
to own any successful technology, because I can turn around and sell
it or use it for higher profit.
3) CMU is privileged because of a long record of excellent work. The
Red Team got the money because of what they have done in the past,
which is perfectly valid to me. Why would anyone invest in a group
that has done little in comparison? They certainly couldn't work
nearly as hard as I've seen the Red Team work. But that isn't even
enough. The course beat everyone. The problem is a fundamental lacking
in the state of the art. High speed dynamics, amazingly accurate maps,
and robust stabilization are all new technologies that CMU has
developed. But there needs to be more.
4) You can't view the race independent from our entrepreneurial free
market system and the nature of government defense spending. Clearly
both will reward success and should be motivators by themselves.
Saying this race is unfair ignore the existing infrastructure which
supports innovation at whatever level gets the job done. It would seem
that at this point, university and corporate research are what get the
job done. Maybe some folks on robot.net can change that. As someone
who will enter the small, growing business mentality of an
entrepreneur very soon, I certainly hope this is the case.
1) The rules where in fact changed to allow entrants to compete that
would not have made it through the QID.
2) I cant speak for everyone, however, the prize money and or possible
contracts resulting from a successful run have absolutely no bearing;
it's the thrill of competing, taking part in the event. I would be
just as enthusiastic about taking part in the event if the reward for
wining was a trophy and a handshake.
3) My comments where NOT intended to detract from the effort put
forward by any of the contestants, especially CMU. I have a lot of
respect for CMU, the students, the facility and everyone else involved;
the point I was trying to make is simple, this problem can not be
resolved by simply throwing money at it; DARPA understands this, hence
the Grand Challenge.
4) Your fourth statement is contradictory, how can you have a "FREE
MARKET" situation when teams are pre excluded from the competition. In
what way do you consider what took place with this event
entrepreneurial? From an outside in perspective this event appeared to
be more closely related to nepotism than a free market.
Carification, posted 16 Mar 2004 at 09:50 UTC by Balthaser »
I doubt u have to tell people about your technology just because u won
that 1 million dollars right.
How you want to tell people of your secrets is up to you. you'll be
under some pressure but i dont think u have to tell them everything...
is that even if the vehicles crapped out in the first mile they still
a)start and move
b)navigate several twists and turns
c)control their speed to do so
all of those being significan hurdles when kludging together
automechanical, computer and sensor technologies. It also shows that it
can be done... not well yet, but it can be done, which is important in
itself. Since it WAS done by teams of all financial backgrounds using
1) Changing the rules made little difference in the end. How many bots
made it 300 yards which didn't pass the QID. Allowing folks to compete
that didn't pass through a presumably less stringent initially pruning
would just mean a lot more failures at the starting line.
2) If the race isn't about money, then what is stopping people from
competing independent of the race? This is a debate about whether the
initial pruning was right or wrong, correct? If people aren't
competing for the prize money, then feel free to continue working, and
I'll shake your hand if/when you win and give you a paper-weight.
3) I believe in challenges like this, especially when you consider how
much work was done for the cost. BUT, there are limits on collective
good will efforts. Someone needs to pay the rent, and it should be
DARPA if you are working on their project full time. Put another way,
as in my original point, DARPA support would ALLOW people to work full
4) There is no contradiction. I am talking about 3 systems.
3-government prizes for challenges.
All three will reward success, which is why our modern age has seen
such amazing amounts of progress. This is like any system of research.
A university does not accept every person that applies, and lets them
all have a chance at success. A university accepts those that seem to
be capable of something good, given a threshold, and goes from there.
It isn't nepotism to limit the contestant field, it's just pragmatic.
BUT, I don't think it would have made that great of a difference
either way, really, as I mentioned above. It WOULD have made the live
satellite feed MUCH more interesting to see more bots fail :)
Lots of ways of interpreting that cliche, however the one I have in
mind is as follows.
If you couldn't enter because they limited the entries (or in my
because I'm one of those dangerous Canadian foreigners ;-), build
'bot that can successfully navigate the course in the alloted time
(and recognize endangered wildlife in the process). Once you've done
so, and your success has been well and officially documented, don't
settle for anything less than a $2M reward!