New quadrupedal robots, based on years of research alongside some amphibian inspiration, could one day crawl and shimmy their way into search-and-rescue operations. As detailed in a new paper recently published in Nature Communications, the robotic trio developed by a team at Colorado State University can swim, walk, and crawl depending on their environments’ obstacles—thanks in large part to lightweight artificial muscles that don’t require heavy onboard power sources.
The new systems, which have been in development since 2017, were designed by a team led by CSU Department of Mechanical Engineering professor Jianguo Zhao, and rely on materials that change rigidity depending on temperature.
“Our embedded morphing scheme uses a lightweight artificial muscle similar to a human muscle, and it contracts when electricity is applied,” Zhao explained in the project’s October 2 announcement. “By embedding these artificial muscles in the spine of the robot or in its skin, we can achieve a variety of shape-types. Altogether, this approach offers a promising path towards developing robots that can navigate and work in difficult environments.”
Aside from the electrical properties, the robots owe their movements in large part to frogs—or, rather, frogs’ multiple life stages. “They start as tadpoles with tails for swimming before developing legs that let them jump, crawl or swim,” Zhao continued. “We take inspiration from those transformations, but achieving animal-like embedded shape morphing in robots remains challenging and is something we hope this work will continue to address.”
Judging from the video montage, it’s easy to see the frog analogy. Depending on its surroundings and terrain, the robots can curve their limbs to “swim,” then adjust them accordingly to scale a rocky hurdle that mimics a shoreline. On dry land, Zhao’s robots can “hop” along by repeatedly rotating their limbs 360 degrees to push forward. A third version of the robot can flatten itself to skitter through small openings, as well as hang onto a ledge to help transition across gaps.
For now, however, the robots require remote control, but future iterations could rely on sensor- and camera-based analysis of their environments for navigation, and even morph as needed to handle their surroundings.