Flash
back to October 8, 2005: Stanford Racing Team's Stanley—a
tweaked VW Touareg with robotic innards—kicks up a cloud of
triumphal dust as it crosses the finish line at the DARPA Grand
Challenge in Primm, Nevada. Stanley was the first of four unmanned
vehicles that day to complete a 132-mile desert course replete with
rocks, treacherous inclines, and cliff-side hairpin turns. Though
Stanford's team grabbed the $2 million top prize, the event
represented a collective victory for autonomous vehicle developers.
After Stanley and its cohorts proved their mettle, self-driving cars
started to seem less like a Jetsons-inspired pipe dream and more like
the automotive wave of the future.

DARPA
is upping the real-world ante with this year's Urban Challenge, held
in Victorville, Calif. The field of 11 robotic finalists won't just
need to steer, turn, and brake successfully to navigate the 60-mile
course; they'll also have to obey traffic laws and signals, negotiate
merges on lane-marked roads, and carry out simulated battlefield
supply missions. The race organizers' immediate objectives are
military—Congress has pledged to replace one-third of its
operational ground fleet with autonomous vehicles by 2015—but if
this year's entrants can pull off bravura performances, the civilian
implications will be enormous. Robotics engineers envision a new
generation of computerized cars that will redefine the term
autopilot, ferrying passengers unassisted from point A to point
B, maintaining ideal speed at all times, and braking for dogs and
bikes faster than any human driver ever could.

The
first vehicle to complete the course will win $2 million, and second-
and third-place finishers will bank $1 million and $500,000
respectively. Flag fall will take place at 8 am Pacific time
tomorrow—stay tuned for up-to-the-minute live coverage of the race.—Elizabeth Svoboda

Blinding rain. Careening traffic. Distracted drivers. There are lots of reasons why car crashes are America's leading cause of accidental death. And one way that most accidents could be prevented: with cars that predict a coming collision-and take action to stop it.