from Nanowerk News:
Held at a former air force base in Victorville, Calif. in late 2007, the DARPA Urban Challenge offered a $3.5 million purse to competitors who could design the fastest and safest vehicles that could traverse a 60-mile urban course in moving traffic in less than six hours. The contestant vehicles were unmanned and had to complete a simulated military supply mission, maneuvering through a mock city environment, avoiding obstacles, merging into moving traffic, navigating traffic circles, and negotiating intersections — all while conforming to California driving rules. Of the 89 international teams that entered the challenge, only six finished in the allotted time.
Wende Zhang of General Motors was part of the team that designed the winning vehicle, which finished with the fastest time — an average speed of approximately 13 miles per hour. The GM team drew upon existing technology already offered in some of their vehicles that can assist in parking or detect lane markers and trigger alarms if the drivers are coming too close to the shoulder of the road. For the DARPA challenge, they developed a more sophisticated package of sensors that included GPS coupled with a camera and a laser-ranging LIDAR system to guide and correct the vehicle’s route through the city. In Baltimore, Zhang will present GM’s patented new methods for detecting lanes and correcting a vehicle’s route, which helped them win the challenge.
Though they won, don’t look for robotic chauffeurs immediately. The technology must prove reliable in many different road, weather and lighting conditions. Still, says Zhang, a commercially-viable autonomous driving product may be available in the next decade.
