Sunday, November 13, 2005

Darpa Grand Challenge

by Max Alsgaard-Miller (Class of '05)

Last month was the second running of the Defense Advanced Research Projects Agency (DARPA) Grand Challenge. The Grand Challenge is a race of autonomous ground robots that must complete a challenging 132 mile off-road course within ten hours. The hope of the Grand Challenge is that the private sector will develop new technology that the military can use without excessive research and development by the military. The first year the challenge was a surprising failure. The prize was one million dollars and the farthest distance was 7 miles. This year's competition, with the prize money raised to two million dollars, was much more exciting. Four robots crossed the finish line within the 10 hour time limit; all of which finished within a 35 minutes block. After all the times were complied it was determined that "Stanley", Stanford's VW Touareg, was the winner.

This year, the qualification was a more rigorous process where robots were placed in five different, very difficult situations, so only realistic competitors would be included in the actual event. These qualification rounds quickly showed where the teams had short comings. Forty robots participated in the qualification round and only twenty-three were selected for the race.

The course was a very challenging 132 miles though parts of California and Nevada. Many elements had to be taken into account when designing a car for the competition. The course was deliberately designed to thwart vehicles that simply wished to use GPS and drive though the course. Many large obstacles were placed throughout the course and sections of the course were in tunnels or on steep passes where the GPS signal weak or nonexistent. Therefore, along with GPS, many of the robots, including "Stanley" utilized laser guidance systems, radar, cameras and active suspension monitoring. The true challenge was getting all these systems work synergistically to predict the best path for the robot flawlessly. Another lesson learned by the teams was, that slow and steady wins the race. Many of the robots that did not complete the course used a strategy that required pushing the fifty MPH speed limit. Most of the robots in the challenge had some sort of error at one point or another during the race but the slower ones could recover from a slight over or under steer while the faster ones would be thrown much farther off course and in some places off of passes. This is clearly shown by the average speeds of the four robots with the best time, 18 MPH.

The development of a truly antonyms vehicle with reasonable speed over long distances is something that the Department of Defense has poured millions of dollars into. The DARPA Grand Challenge was an excellent enlistment of non-federal parties to do the research and development for the government. Now that it has been completed we will just have to wait and see what they come up with next.

References:

Links to this post:

Create a Link

<< Home