DARPA Grand Challenge Finalists Announced 129
Xerotope writes "DARPA announced today the 23 finalists[pdf warning] of the DARPA Grand Challenge at the closing ceremonies of the National Qualifying Event. Carnegie Mellon University's Red Team will start on Saturday with the first and third positions, with 'H1ghlander' taking the pole position and 'Sandstorm' following 10 minutes later. Stanford's 'Stanley' will start second. Of the 43 semi-finalists, 23 robots managed to finish the 2.2 mile course at least once. 5 robots (Stanford, Red Team, Red Team Too, Axion Racing, and Team Teramax) completed all of their runs. CMU's 'H1ghlander' and 'Sandstorm' finished the four runs with an average time of 10 minutes, 20 seconds each. Stanford's Stanley average time was 10:43."
Ghost Rider (Score:3, Informative)
Go ENSCO! (Score:2, Informative)
www.teamensco.com
More information and Video coverage here (Score:5, Informative)
http://www.cartv.com.nyud.net:8090/content/resear
NQE final paper:
http://www.darpa.mil/grandchallenge/NQEfinal1.pdf [darpa.mil]
And more announcements can be found on:
http://www.grandchallenge.org/ [grandchallenge.org]
Also, a good summary of things that have been happening can be found in the discussion forum:
https://dtsn.darpa.mil/grandc/forum/topic.asp?top
=====
A post by Espina reads:
Hi!
well if the summary isn't going to explain it... (Score:5, Informative)
"The DARPA Grand Challenge is an unprecedented government effort to accelerate research and development in autonomous ground vehicles to help save American lives on the battlefield. DARPA will award $2 million to the autonomous (robotic) ground vehicle that can successfully navigate a challenging desert course of approximately 150 miles the fastest (in less than 10 hours). The vehicles must find and follow a prescribed course route, avoid obstacles, and negotiate turns, all while travelling at militarily-relevant rates of speed. The ground vehicles are fully autonomous - not remote-controlled."
Re:More information and Video coverage here (Score:3, Informative)
http://www.tgdaily.com/2005/10/06/darpa2005_featu
"Blue Team" runs self-righting motorcycle at darpa Grand Challenge
http://www.tgdaily.com/2005/10/03/darpa2005_featu
DARPA Grand Challenge update #3: Interview with Team Cornell
http://www.tgdaily.com/2005/10/02/darpa_cornell_i
DARPA Grand Challenge update #2: A chat with Team Mojavaton
http://www.tgdaily.com/2005/10/02/darpa_grand_cha
DARPA Grand Challenge Update #1: Qualification Day 1 results
http://www.tgdaily.com/2005/10/02/darpa_grand_cha
Gearing up for the DARPA Grand Challenge:
http://www.tgdaily.com/2005/10/01/darpa_grand_cha
Re:Still too slow (Score:4, Informative)
At the qualifications, there were mandatory speed limits imposed in most (if not all) the areas. In the RDDF file, there are the GPS coordinates and a speed limit number. For example, the straight away was 40mph while some of the obstacle strewn areas was 5 mph. The vehicles are capable of going faster and in fact a couple vehicles maxed out the 40 mph on the straight away.
DARPA officials at the media press conference on Wednesday said that if they stick to the race speed limits, then they will finish in about 6 1/2 to 7 hours. In the real race, there are hard speed limits and then there are suggested limits (which a team can break). The suggested speed limits are in low obstacle areas, but are suggeseted so that the chase vehicle doesn't lose sight of the robot. Remember that this race is being held in the desert, so the dust kicked up could obscure it from view.
Quirky fact, the BLM (Bureau of Land Management) mandates 25 mph or less in desert tortoise areas. You gotta love beauracracy.
Some of this is explained in my article on tgdaily.com
http://www.tgdaily.com/2005/10/06/darpa2005_featu
Re:Still too slow (Score:3, Informative)
At least three of the vehicles that qualified for the race have performed 170 mile runs at race-success pace on mixed road/off-road courses that should simulate race conditions very closely. Many of the vehicles have code in place that allows top speeds over forty miles per hour when long, straight, smooth sections of course are detected.
I know that the CMU teams pre-plan their runs in the two hour period between receipt of the course waypoint file and the beginning of the race. They will not load a plan designed to execute in more than ten hours and, given the quality of the competition this year, I betcha they'll be aiming at eight.
A hot issue this year that I haven't seen discussed on public forums is intent-to-pass. My understanding is that DARPA will force a vehicle into 'pause' mode if it is being overtaken by a faster competitor. I'm willing to bet there will be some post-race howling around that dynamic.