Sep 15, 2011 - Claremont, CA -

Imagine a robot that could navigate its environment much like a human, receiving visual cues and comparing them to stored images to interpret its location at any given time. Such a robot could prove useful for search and rescue operations, transportation, military applications and helping the visually impaired.

This summer, Lilian de Greef '12, Brad Jensen '13, Kimberly Sheely '12, Malen Sok '13 (Cal Poly Pomona) and Nick Berezny (Cal Poly Pomona) worked with faculty advisor Zachary Dodds, associate professor of computer science, to test various computer-vision-based algorithms and machine-learning methods that could help a robot autonomously navigate its environment.

Working with an ARDrone quad-rotor helicopter, equipped with two webcams, the team explored two methods of localization. The first captured and stored images of a designated physical space, while the second employed April Tags to mark specific physical locations. The team then developed software that compared the current images registered on the helicopters' webcams to either the stored images or tag designations. Using these systems, the helicopter could autonomously determine its current location and navigate to other, specified areas.

The project was part of Harvey Mudd College's Summer Research Program, which engages students in 10 weeks of full-time research. More than 170 HMC students pursued research projects this summer alongside 45 faculty members in biology, chemistry, computer science, engineering, mathematics and physics.