A team of eight undergraduates and their robot Mabel the Mobile Table won the American Association for Artificial Intelligence 2002 Mobile Robotics competition in July. The group was the only team at the Edmonton, Alberta conference to be made up completely of undergraduate students.
The team was spearheaded by senior Jonathon Schmid, who attended the conference last year to gauge the competition and plan a strategy for this year. With the help of computer science professor and adviser Chris Brown, a team of 11 students did independent research on the robot during the fall of 2001, and nine worked on Mabel in the spring of 2002. The summer team of competitors was comprised of juniors Mike Isman and Tom Kollar and seniors Eric Meisner, Jenine Turner, Tori Sweetser, Ben Atwood, Dave Feil-Seifer and Schmid.
Several students were paid to work on the robot this summer through Research for Undergraduates, a grant from the National Science Foundation. Feil-Seifer, the team’s business manager, applied for travel grants from AAAI, which covered the conference registration, housing and some of the travel expenses.
The team’s dedication was tested during the two week trip to the conference. They traveled the 4,600 miles to Edmonton in Schmid’s 1993 Toyota Previa, staying at friends’ houses along the way. “The cross-continent road trip was perhaps one of the most fun parts of the experience,” Schmid said.
“Our kids worked very hard, with little assistance from faculty and grads,” said Marty Guenther, secretary for the computer science department. “The kids deserve a huge pat on the back for their work.”
Meisner stated that the best part of the experience was “learning from the other team members and finding our own solutions to problems we encountered,” while Schmid said that, “the best part was proving that artificial intelligence has reached a point where a group of motivated individuals can build a real robotic system capable of accomplishing useful tasks in the real world.”
The conference in Alberta was Mabel’s competition debut. The goal was to design a robot that could serve food. Mabel was not only designed to serve food but could also understand speech, read nametags, track the speaker’s face and navigate through a crowd. The robot used a speech recognition microphone and face tracking with a pan-tilt-zoom camera to interact with humans. Sonar-based obstacle avoidance was used to help the robot navigate a crowded room.
“It was exciting to know that we could build a state of the art system without a grad school budget and doing our own research,” said Schmid.
Mabel won the Host Competition at the AAAI conference and was awarded a first-place certificate. Future plans for the robot still remain uncertain, although several of the students would like to continue working on her throughout the year and possibly enter next summer’s AAAI conference in Mexico. You can catch Mabel in action as part of Meliora Weekend.
DeSantis can be reached at email@example.com.