Yale researchers create Amphibious Robotic Turtle
NEW HAVEN — What swims in water and walks on land, and has a shell and morphing limbs?
A turtle. But instead of bone, flesh and cartilage, this creature consists of hardware, plastic, silicone and rubber.
ART, an Amphibious Robotic Turtle, was created by a team of Yale engineers.
Its lair is a small room inside Yale University’s Mason Laboratory on Hillhouse Avenue. A sign on the door to the room reads, ‘WARNING: This property is protected by a highly trained turtle.’
The creation of ART was headed by Rebecca Kramer-Bottiglio, an associate professor of mechanical engineering and materials science at Yale.
Her lab focuses on “studying the fundamental principles that guide how and when a robot should adapt its morphology and behavior when transitioning between tasks or environments,” Kramer-Bottiglio, who was not available for an interview by deadline, said in an email.
In 2017, Kramer-Bottiglio wrote a grant for the project, whose goal was to create a morphing robotic limb, according to Robert Baines, a PhD student in Kramer-Bottiglio’s department who was a first author on the research paper about ART.
“I came into the lab, and I worked on that morphing robotic limb, and as time went on we wanted to incorporate it into an entire robotic system,” Baines said.
According to a release from Yale, other authors of the study affiliated with the university included Sree Kalyan Patiballa, Joran Booth, Luis Ramirez, Thomas Sipple, and Andonny Garcia; Frank Fish from West Chester University also contributed.
The team’s goal was initially to create a limb take the shape of a flipper to efficiently swim in the water and shift to a semi-circular load-bearing shape to walk on land, Baines said, equating the difference to the limbs of a sea turtle to those of a tortoise.
“To design amphibious robots is in itself extremely challenging,” said Patiballa, another one of ART’s creators.
Also a first author on the corresponding research paper, Patiballa completed his postdoctoral research working on the project and is now an assistant professor of mechanical engineering at the University of Alabama.
According to Patiballa, though there have been many amphibious robots, most have used two different propulsion technologies – wheels to move around on land, for example, and flippers to navigate the water.
In ART, the research team sought to combine two morphologies inspired by biological adaptations found in nature, Patiballa said.
A sea turtle’s flippers are efficient on water but not on land, he said, while other turtles’ semi-circular limbs pose the opposite problem.
The team’s approach is a “new design paradigm which we thought would pave the way to the next class of adaptive robots,” Patiballa said.
To create limbs that could change shape, the researchers used a type of plastic that becomes soft when heated.
By applying heat and pressure, Patiballa said, the limb becomes flexible and changes shape. It then holds that shape after it cools and stiffens, he said.
Waterproofing the robot proved one of the project’s biggest challenges, Patiballa said.
According to Baines, another challenge has been helping ART navigate the physical phenomena found at the boundary between water and land, such as pebbles, sticks, sand and waves.
“The outdoors was our lab space,” said Ramirez, a PhD student in mechanical engineering at Yale who worked on the project. “It was pretty cool seeing robotics in real-world settings.”
Researchers tested ART at East Rock Park’s canoe launch. The project’s next step will focus on the land-to-water transition zone, according to Baines.
“The overall question we want to explore is when should the robot change its shape, when should it change its movement patterns, and then how it should make that change as well,” he said.
Right now, ART is tethered to a power cord, and a human operator controls when its limbs change form, Baines said.
But eventually the team hopes to create an autonomous robot whose limbs morph in response to environmental stimuli, Baines said.
Such an accomplishment could have real-world applications, including environmental monitoring that provides scientists with continuous data streams, Baines said.
“We need some kind of a robot that can monitor and even eventually clean without disrupting its surrounding environment,” said Patiballa.
An autonomous amphibious robot could, for example, be deployed on shorelines and in other aquatic-terrestrial environments to catalog biodiversity, Baines said.
“The notion of the adaptive shape-changing robot is going to be something of the future,” he said.