By Sandy Field
Keith Buffinton, professor and co-chair of mechanical engineering at Bucknell, has a longstanding interest in the development of legged robots. His work involving the development of energy-efficient, bipedal walking robots has received attention from the National Science Foundation (NSF) and the U.S. Department of Defense. Last June, Buffinton and collaborators Ken Waldron, a Stanford University professor, and Alex Perkins ’05, a Ph.D. student at Stanford, received NSF funding for a project to improve existing bipedal robots. They believe current robots, designed for stability, would be more useful and energy efficient if they were capable of walking more like humans. Humans can negotiate difficult terrain by maneuvering around and over obstacles and can stop and start quickly in response to external stimuli. “The goal of our project is to model, analyze, and develop a robot that can execute truly dynamic movements such as running, jumping, turning, starting rapidly and stopping suddenly,” explains Buffinton.
The work has recently gained national attention as part of an $800,000 federal research grant – part of a Defense Department Appropriations Bill for 2008. Buffinton and Bucknell collaborator Steve Shooter, professor of mechanical engineering, along with researchers from the Institute for Human and Machine Cognition, will work to develop bipedal robots with potential military and law enforcement capabilities. Bucknell and Congressman Chris Carney, who worked with Bucknell to secure the funding, announced its availability in November 2007. “We will travel to visit IHMC researchers in mid-January to discuss the goals of this project,” says Buffinton. “We are very excited about the prospect of working with the researchers at IHMC and hope that this project provides a basis for ongoing collaborative work.”
In addition to the bipedal-robots projects, Buffinton recently returned from a sabbatical in Switzerland, where he worked on a project developing quadrupedal robots. He is working in the laboratory of Roland Siegwart of the Autonomous Systems Laboratory at the Institute for Robotics and Intelligent Systems at the Swiss Federal Institute of Technology. “The goal of the quadruped project is the design and construction of an energy-efficient, four-legged robot capable of various dynamic motions, including fast walking, trotting, bounding, and galloping,” says Buffinton.
How is this different from the bipedal work? “Many elements of the modeling and dynamics of bipeds and quadrupeds are similar, as are the drive mechanisms,” explains Buffinton. “The motions are different and the approaches to producing them also differ, but experience with one can definitely be transferred to the other.” Buffinton hopes to improve existing designs for quadrupedal robots by taking advantage of the natural momentum-induced swinging of the legs to achieve significant energy savings. “Our hope is to develop a new form of quadruped that uses what are known as “series-elastic actuators” as well as new forms of control algorithms that naturally lead to more efficient movements,” he says. “We can’t wait to see how they run.”