Vehicle engineered with a mind of its own

It’s official: “Warthog” is the rookie of the year, at least when it comes to robots.

To be precise, it’s Warthog’s creators—a team of UD graduate and undergraduate computer science and engineering students—who captured rookie-of-the-year honors at the 15th annual Intelligent Ground Vehicle Competition held this June in Detroit. It was the University’s first appearance at the event, which offers its competitors cutting-edge design experience.

Warthog, based on a rugged four-wheel platform manufactured by Segway Inc., is equipped with a laser range finder, a color camera and differential global positioning system.

The robot was built by a team of 15 students, who were advised by Christopher Rasmussen, assistant professor of computer and information sciences. The team also programmed the robot to meet two separate challenges involving autonomous obstacle avoidance and navigation in an outdoor environment.

“We only finished building the robot a little more than a month before the contest—much later than we had planned,” Rasmussen says. “Unfortunately, within the first few days of outdoor testing in late April, the motors had a major breakdown, and as soon as they were repaired by Segway, they broke again with only 10 days to go. It took a lot of toil and luck to even make it to Detroit with a functioning robot, so I’m very gratified by our results.”

Five of the students—Sherol Chen, Amit Hetawal and Ezra Kissel, all graduate students in the Department of Computer and Information Sciences; Donald Scott, AS ’08, a computer and information sciences major; and Patrick Griffith, EG ’09—represented UD at the competition.

They gave an oral design presentation to the judges and shepherded the robot through a safety inspection to qualify it for the main competition.

“The students worked hard throughout the semester, and they really came through in the last week when we thought all might be lost,” Rasmussen says.

“Seeing Warthog managing to navigate around barrels and fences in the contest just days after being a pile of parts on the lab floor was a great moment. We learned a lot at the competition, and now that the robot’s working, we will be a team to reckon with next year.”

The competition, which this year drew teams from nearly 40 institutions, including Virginia Tech, Georgia Tech, the University of Texas and the University of Minnesota, was a valuable learning experience, participants say.

“Just being there taught me a lot personally, so much so that I think we will be a legitimate threat to win it next year,” Griffith says. “We had to deal with several hardware issues that severely changed our approach, but we were still able to make a successful appearance that earned us the rookie of the year award.”
Chen says the process of seeing the Warthog project come together was interesting in itself.

“It was definitely worthwhile to see how such large projects are managed,” she says. “Once we got to Michigan, I found the competing teams to be very collegial. The professionals running the competition were supportive, encouraging and eager to lead the cutting edge of robotics.”

Scott says the competition was an opportunity for the students to see all of their ideas at work in a practical application and an experience that is almost impossible to gain in the classroom alone.

“It was a chance to work together with my teammates and faculty adviser toward a common objective and to compare our work to that of other teams from many other universities,” he says. “It was an invaluable learning experience. I became familiar with so many new and useful algorithms and software systems.

Consequently, I am now more than ever excited about doing similar research in graduate school.”

Warthog was given the rest of the summer off, but preparations soon will be under way for next year’s competition, where Rasmussen says the UD robot “figures to be among the top contenders.”

Rasmussen, a hiking enthusiast who became interested in artificial intelligence while an undergraduate at Harvard University, conducts groundbreaking research in computer vision for field robotics, specifically methods of helping mobile robots find and follow trails.

“I’m fascinated by the quest to get computers to behave like people, to exhibit human-level intelligence at various tasks,” he says.

For more information about the Intelligent Ground Vehicle Competition, visit [www.igvc.org].