In a brief but spectacular racing career, Stanley beat twenty-two other robot vehicles for the $2 million prize in the Grand Challenge, held in October 2005 on a demanding 132-mile desert course near Las Vegas, Nevada. The goal of the race, sponsored by the Defense Advanced Research Projects Agency (DARPA), was to stimulate invention for a future fleet of driverless military ground vehicles. Congress funded the competition to support its directive that one-third of U.S. military ground vehicles be unmanned by 2015.
Stanley represents a promising research direction in artificial intelligence, or machine thinking. Through sophisticated programs in onboard computers, the vehicle decides how to navigate mapped terrain and unmapped obstacles in real time. It integrates a course map expressed in about 3,000 points of latitude and longitude, stored memory of past experiences, and new information about the road ahead gathered from roof-mounted laser sensors, video cameras, radar and GPS receivers.
Behind Stanley’s driverless accomplishment is the work of nearly 100 people at Stanford University and Volkswagen’s Electronics Research Laboratory (ERL), both in Palo Alto, California.
DARPA’s Grand Challenge of 2005 pitted autonomous vehicles against each other and a ten-hour limit on a punishing dirt course with steep cliffs, sharp turns, and countless obstacles. Only Stanley and four other competitors finished the course. The race’s experimental robots—all sponsored by businesses, universities and individuals—emerged from research for military purposes and demonstrated the feasibility of self-navigating vehicles.
Like the impact of integrated circuits, the Internet, and other technologies with strong military connections, the impact of the robot race is likely to be felt in other areas of American life, especially automotive safety.