You have commented 339 times on Rantburg.

Your Name
Your e-mail (optional)
Website (optional)
My Original Nic        Pic-a-Nic        Sorry. Comments have been closed on this article.
Bold Italic Underline Strike Bullet Blockquote Small Big Link Squish Foto Photo
Home Front: Tech
The “Grand Challenge” Robot Vehicle Race
2005-09-14
September 14, 2005: DARPA (Defense Advanced Research Projects Agency) is holding it’s second “Grand Challenge” race, on October 8th, for developers of UGV (unmanned ground vehicles.) DARPA is offering a two million dollar prize for the first UGV to complete the 280 kilometer course, in less than ten hours. The race, first held last year, has not yet been won.
The actual course is not announced until two hours before the race begins, but the finalists are told what general area to be in several weeks before the race. This is done so participants cannot electronically survey the course, and simply program an UGV on how to best navigate it. The purpose of the Grand Challenge is to motivate vehicle manufacturers to create the technology the Department of Defense wants for large UGVs that can carry supplies and equipment without the need for drivers. In theory, this capability has been possible for several years. But no one has actually done it yet. Research on this subject has been going on for decades.

DARPA was correct in assuming that the Grand Challenge would get the creative juices going, and cost the government little. Participants get no money from the government, unless they win. And if they do win, military vehicle manufacturers will be very interested in buying the technology that made it possible. Naturally, many military vehicle manufacturers have entries in the Grand Challenge. Oshkosh Truck, for example, has partnered with electronics manufacturer Rockwell Collins, and artificial intelligence experts from the University of Parma, Italy, to produce TerraMax. This vehicle has, in the last few months, consistently moved over cross-country routes at 31 kilometers an hour (and sometimes as fast as 56 kilometers an hour), meaning it’s a contender for winning the prize. TerraMax has sensors and an onboard computer that monitor the pitch and roll of the vehicle, as well as how close it is to the edge of the road, or any obstacles, and adjusts steering and speed accordingly. TerraMax has been so fast at times that the SUV “chase car” could not keep up. That’s because TerraMax has no people on board, and a better suspension system for dealing with the rough terrain. The TerraMax vehicle is basically a six wheeled Oshkosh “Medium Tactical Vehicle Replacement” (MTVR), and weighs over 14 tons, and is the largest race participant.

In the first race last year, no vehicle completed the course. In fact, the best any vehicle could do was 23 kilometers (about ten percent of the entire course.) This year, several participants feel they can not only finish, but do it quickly enough to win the prize.
Posted by:Steve

#2  Attention Attention..calling all Ships..calling all Ships. <)

/so solly Moosey. ;)
Posted by: Spiting Whaviper5458   2005-09-14 14:46  

#1  The entire "Battlebots" philosophy will do more to stimulate innovation in design than just about any other technique. Darwinistic design evolution is even being done at the software level to see if virtual organisms can be created that mimic real ones. It works.

I might suggest thinking outside the box, however. Right now, everyone probably thinks of having a complete "brain" in the vehicle, using satellite GPS for guidance. However, practically speaking, this only gives the vehicle a two-dimensional view of the terrain. This would be the way a human driver would have to do it, but the robotic guidance does not have that inherent limitation.

That is, if they permitted low and/or medium altitude guidance from a non-entrant aircraft, it could give them an essential angular 3rd dimension to evaluate the terrain with. Either a government UAV or military aircraft, providing equal info feeds to all participants.

Right now, they expect too much from their integrated guidance. Computer brains are hard-pressed to handle abstracts like perspective, instant evaluation of soil type, objective and subjective navigation, etc. So they rely too much on trying to "brute force" their way through innumerable obstacles.
Posted by: Anonymoose   2005-09-14 12:22  

00:00