Using dance to design robotics control systems
March 27, 2016
Amy LaViers, assistant professor of mechanical science and engineering at the University,AG is using this connection in her upcoming research project that aims to create a multi-platform robotics control system inspired by the way humans move.
The project, titled “Choreography of Embodied, Platform-invariant Motion Primitives,” recently won a Young Faculty Award from the Defense Advanced Research Projects Agency, which is providing funding for the project.AG
With the help of two graduate students, Umer Huzaifa and Anum Jang Sher,AG LaViers hopes to create a control system that is more expressive and intuitive for its human operator.
LaViers said addressing these goals will require the incorporation of Laban/Bartenieff Movement Studies, a classification for describing movement that is used by choreographers and has been an inspiration for robotics research.AG
Think of a dance troupe. Each individual person is unique, yet the dancers all appear to move in unison.
“That language out there exists. People know how to use it,” LaViers said about dancing. “We should leverage that intelligence in our robotic algorithms.”
LaViers said this project will be a deeper exploration into how dance and robotics interact. For example, their grant will fund workshops with certified movement analysts who are proficient in Laban movement analysis, and LaViers herself will become a certified movement analyst.
The control system they will create could be used in a variety of real-world applications, such as manufacturing or prosthetic devices. But the link to improving national security is what led to the grant from the Defense Advanced Research Projects Agency.AG
LaViers recently visited the United States Military Academy at West Point, where they use a robot called a PackBot to perform surveillance and dismantle suspected explosive devicesAG. The PackBot is intended to perform tasks so no human life is at risk.AG
But the capability of the PackBot’s current control system is limited. For example, the robot cannot simultaneously pull in its robotic arm and move backward.
“That’s the exact kind of thing that is a more complex, more expressive behavior that this proposed methodology that we’re developing can address,” LaViers said.
Kyung Yun Choi, AGa graduate student researching soft robotics at the University but who is not involved with the research, said this project could build a new connection between humans and robots.
“By investigating the human’s behavior in order to build a robot to reflect our general motions in natural way and have it express a variety of expressions, I hope that we could learn more about ourselves as well as new aspects of developing a robot,” Choi wrote in an email.
Andy Borum, AGa graduate student in aerospace engineering who is also not involved with the research, said LaViers’ work could make a difference in medical procedures, manufacturing, and search and rescue missions.
“Having robots work collaboratively with people to perform complex tasks is a challenging problem,” Borum said. “(LaViers’) work would allow users, particularly nontechnical users, to control the qualitative behavior of robotic systems, allowing robots to work safely and more efficiently alongside humans.”
But LaViers acknowledged that this kind of research has its limits when it comes to imitating human movements.
“(Robots are) very far away from the capabilities of movement that humans have,” LaViers said.
She said this research may also address the ethical concerns that people might have about robotics in general by engaging with this community and helping research evolve.
“I see robots as machines, and I see machines as important tools that help people,” LaViers said. “We need to design (robots) very carefully … that’s a really important job. We need lots of good people working on that problem.”