DARPA Announces Grand Challenge 2005

An anonymous reader writes "The Grand Challenge 2005 Date has been announced for October 8, 2005. Check out DARPA's official webpage for details. Already several teams from last year are gearing up: Carnegie Mellon Red Team, D.A.D., and Cal Tech. Also, several new teams are entering, among them Stanford, and Florida Tech. Should be a very interesting Challenge next year!"

Comments

we can only hope

[bigben7187]

maybe this time someone will actually make it more than a few miles. Actually i expect someone will finish the whole race, but i do not know if they will finish it within the time requirements.

Prize increased to $2M

[MarkGriz]

The team that completes the Grand Challenge 2005 route within a specified time will receive a cash prize of $2 million, an increase from the $1 million prize offered at last year?s event.

In the first DARPA Grand Challenge, held on a desert route from California to Nevada, 15 teams from a field of 106 applicants progressed to the final event

Why increase the prize to $2M? If the goal of the challenge is to develop an autonomous vehicle, why not use the extra $1M as a grant to fund the top 15 teams from the last challenge.

Nitpicks

[GillBates0]

Anthony J. Tether, DARPA director, noted: "This event is a challenge for American ingenuity. It brings together individuals and organizations from the research and development community, industry, Government, the Armed Services, academia, professional societies, and from the ranks of students, backyard inventors, and automotive enthusiasts.

Are non-American citizens allowed to participate? I tried looking at the Rules Page but it's not up yet. I don't recall if there was a stipulation which restricted participants to American citizens.

Given DARPA's great R&D track record in the past (Internet and what not), I would've liked to participate in the contest *purely* from a scientific curiousity point of view - and I bet a lot of nerds all over the world would like to overlook the fact that the contest is sponsored by a military agency (prize not withstanding - since it's US taxpayer money). Just as long as DARPA lives up to it's name and does not morph into OARPA - it's happened way too many times in the past.

Incidentally, the link to the official page is incorrect on that page. The site linked to in the article seems to be just a mirror of the darpa.mil site, however.

Re:I still don't understand why they don't

[Marxist Hacker 42]

And what will you do when this thing crashes at 60 mph into one of the barriers you haven't designed it to detect?

Who's detecting anything? Go ahead and crash into the boulder or whatever- design the robot to bounce, use an accelerometer to detect the bounce. 60MPH bumpers aren't exactly impossible, you know.

Collision avoidance without physical contact is one of the biggest challenges these teams face.

Is that in the rules that they must have no physical contact? The website I saw deleted the rules for 2004 and hasn't posted the rules for 2005- but I can't imagine why they'd care if a combat-bot moving through enemy terroitory demolishes a few houses along the way, or crashes into them, detecting them as a barrier, and moves off in a different random direction.

Also, the obstacles aren't necessarily these big, obvious concrete barriers dumped in the middle of the desert. How would your robot deal with encountering a lake, or coming to the edge of a cliff?

Well, the lake was in the original, that's why you'd want the robot to be amphibious as well as reasonably indestructible. But I grant you the cliff- I guess it depends whether the cliff is on it's internal map of the area or not. Anyway, that's one way to detect the cliff and even find roads/bridges to cross a canyon. Another way to deal with it is with a range finder angled towards the ground- if it's suddenly WAY steeper in front of the robot than the robot was expecting, turn around and find another way.

Or, alternatively, as long as you're making your electric-drive, 60MPH, invertable robot's body out of black-box strength steel to begin with, don't worry about the cliffs. A drop of a mile or so won't hurt it, and it will find it's way out of the canyon eventually.....

wheels vs legs

[HybridJeff]

Were all the entries last year wheel based? Id think when it comes to maneuverability and obstacle avoidance, a walking robot would do the job better. Of course you couldnt have one of those lame assed ASIMO type bots. Do damn slow and clunkey. Just make somthing with at least 4 legs, and you've fixed that whole falling over problem.

Thats what Id try anyway. I'd fail miserably, but wouldnt building a mechanical cheetah b e lots of fun.

Lessons Learned

[texbot]

As a member of Team Caltech I can say that next year's race going to be much more exciting for a couple of reasons.

First off, returning teams already have a foundation to build upon, both in existing hardware and in technical experience. Writing software for a race like this is almost impossible to do without testing it on a working vehicle (i.e. testing in simulation only works out major problems, but does not translate well to desert racing). For example, Caltech spent 6 months getting actuators and drivers to work well enough to hand over the vehicle to the software team. As soon as that happened, we noticed several problems interfacing the actuators to the software (e.g. updating actuator positions too fast locked them up and made them stop responding). Ultimately our vehicle was not even waypoint following accurately until late February. Most teams were in the same state we were in - racing the clock, plagued by bad hardware (sensors and actuators) and inexperience. BUT we were very close to being very good. If they re-ran the race mid-summer this year the results from all of the teams would be very different. Looking to next year, teams have working vehicles which means 1 full year of onsite testing instead of 2 months.

Another thing that is interesting about the next race is the timing relative to academic calendars. A lot of teams are university driven and it was very difficult for students to devote enough time to the project while still handling their school requirements (definitely true with a Caltech workload). The next race is at the very end of the summer which means that a crew can work on the vehicle full time for three months before the next race.

Whether or not someone wins the next race is entirely up to DARPA. By next year there will be 5+ teams that could navegate last-race's course in

Anyway, good luck to all teams...especially new teams - you have quite a hurdle in front of you. See you in 1 year.