Illinois researcher to help design robots that move like living creatures
In the 1960s, street dancers popularized “The Robot,” a dance that describes the stiff, mechanical motion of the artificial beings. But what if robots weren’t so, well, robotic? What if they could jimmy and jive just like humans? That is the fundamental idea behind a recently funded project involving CSL professor Yuliy Baryshnikov, who seeks to help create a robot whose movements more closely mimic that of living creatures.
The project, "Science of Embodied Innovation, Learning and Control," is a joint effort between Baryshnikov and researchers at the University of Pennsylvania, University of California Berkeley, and John Hopkins University to use the natural world, combined with engineering, to design a smoother moving robot.
“A lot of people ask what the right way to design robotics is and what the limits are, but very few people act on it,” said Baryshnikov, an electrical and computer engineering professor at Illinois. “I personally am very interested in looking at it. We have a multi-university team that combines people who work with robotics and animals.”
Each member of the team and their specific expertise will add a crucial component to the development. A material scientist will look at how different animal tissues and joints work together and pass along information, examining the interacting parts and how they can collectively achieve tasks. An animal behavioral psychologist will analyze how animals reason and mentally process prior to movement and apply those findings to add natural flow to a robot’s movement. A roboticist will take all of these findings into consideration when building the final product. Baryshnikov’s job throughout the project is to help his team members set mathematical limits.
“My role as a mathematician is to help everybody,” Baryshnikov said. “I’m looking at the fundamental limits of what can be done. You have to draw some physical realities.”
The limitations that Baryshnikov is helping determine include how much energy can be stored, the highest and lowest amounts of energy that can be created, and temperature thresholds among others. The animal observation will also help Baryshnikov determine what nature’s limits are and what the ceiling may be for the group’s designs.
In addition to a more agile robot, the team also hopes to produce new algorithms for the field of robotics.
“Current robots have trouble walking, and we want a robot that can do parkour,” laughs Baryshnikov. “The different members of our team will be able to design new experiments that make new algorithms and produce new theory. There will be quite a diverse output.”
The project is funded by the Defense Advanced Research Projects Agency (DARPA) at $750,000 for five years. The project fits under DARPA’s ‘Expand the Technological Frontier’ objective which is focused on overcoming scientific barriers and applying the new capabilities to national security needs.