Duke Radar May Give Red Team Competitive Edge in DARPA Grand Challenge Race
Duke University engineering students have designed an onboard radar system to give Red Team vehicles a competitive edge in the upcoming DARPA Grand Challenge race. In that contest, vehicles must run across a desert entirely self-guided without human intervention.
The Red Team is an alliance of students, corporations and volunteers led by the Robotics Institute of Carnegie Mellon University in Pittsburgh. The team is developing two modified Hummers to run in the competition this fall sponsored by the Defense Advanced Research Projects Agency (DARPA).
The race route could be as long as 175 miles over California desert terrain featuring natural and man-made obstacles. The exact route will not be revealed until two hours before the event begins, and the course must be completed within 10 hours. Using radar, the Red Team vehicles are expected to detect obstacles as far away as 180 feet, enabling navigational systems to plan the safest, fastest route.
The radar system, called DROID for Duke Radar Obstacle Identification Device, recently passed performance tests and officially earned a place on the Red Team's two vehicles -- Sandstorm and Highlander.
Until now, radar has never been successfully adapted for off-road vehicle use and the Duke project was considered a gamble, said the Duke engineering students. Although both Red Team vehicles were selected as semifinalists for the race in early June, the radar system was still being evaluated. The Duke students had to prove their technology worked well by mid-June or face being dropped from the Red Team project.
"Our system performed so well that now we're considered a major asset," said Josh Johnston, who graduated from Duke's Pratt School of Engineering in May with a degree in electrical engineering/mechanical engineering. Johnston, from Boise, Idaho, and Jason Ziglar, from Tampa, Fla., a May biomedical engineering/electrical engineering graduate from Duke, continue to lead the Duke team. Johnston is now a graduate student at Carnegie Mellon University and Ziglar is now a Red Team employee.
"The real advantage of our radar system is what is being called 'horizon sensing,' Johnston said. "We are able to look out as far as sixty meters -- almost double the distance of any other sensor system on the vehicles. Without radar, the vehicles can't detect obstacles soon enough to avoid them at high speed. With radar, we've safely navigated at thirty-five miles per hour."
The DROID can detect obstacles as small as 1 foot tall and doesn't get confused by multiple objects, developers said. Additionally, radar is well suited for the desert race because it penetrates smoke and dust, which also are common to the battlefield environment. DARPA's ultimate goal in sponsoring the race is to develop technologies that will save the lives of soldiers on the battlefield.
"This radar work is very tricky, very cool, and the success is very big," said William "Red" Whittaker, robotics professor at Carnegie Mellon University and leader of the Red Team project. "The ability to see well, see far, and see through dust enables higher sustained speed and that adds up to higher performance and lower elapsed time over a long race."
"Duke's radar work is not only significant from a technological standpoint; it improves the autonomous capability of the vehicle, and that is whole point of the race," said Duke's sponsor, Clint Kelly, senior vice president for Advanced Technology Development at Science Applications International Corp (SAIC). "It is tremendously exciting to see a map develop as radar picks out objects on the horizon. As the vehicle gets closer, other sensor systems confirm the objects are there and overlay that information on top of the map display."
Duke students have worked on the system for more than nine months and designed the DROID to be installed on the front of a vehicle for a 180-degree field of view. The team spent months analyzing different obstacles -- fence posts, barbed-wire barricades and natural hazards such as rocks, ravines and trees -- and identifying distinguishing features to create "obstacle filters." Those filters became a catalog for the computer program to consult when trying to identify and classify an object detected with radar. With the filters working well to identify objects, the team then integrated the radar system with the vehicle navigation and decision systems.
"In nine months our DROID team has taken a long-shot technology and delivered a significant advancement in sensor capability to the Red Team," said Robert Kielb, senior research scientist in Duke's mechanical engineering department and faculty adviser for the DROID project.
DARPA announced on June 6, 2005, that 40 semifinalist vehicle teams were advancing to the next stage of the competition. The 40 teams will compete head-to-head in the National Qualification Event at the California Speedway in Fontana, Sept. 27 to Oct. 5. Twenty of the 40 teams will be selected to compete for the $2 million prize in the Grand Challenge event, scheduled for Oct. 8.
Key members of the Duke project include Johnston, Ziglar, Aaron Mosher from Boeing Co., and Danielle Hyatt, a junior in electrical and computer engineering at Duke. The DROID project is one of three projects supported by the Duke Robotics Club.
Red Team Members include students, volunteers and professionals from corporate sponsors, including Boeing, Science Applications International Corp., Caterpillar, Intel, AM General, TTTech, Applanix, HD Systems, KVH, Snap On, Chip Ganassi Racing, Google, CM Labs, HMR Magazine and Wired Magazine.