Robotic Fish Swims as Deftly as the Real Thing
Before robots go mainstream, there are two major motion-related challenges they will need to surmount: navigating complex environments, perhaps by aping nature’s solutions to the same challenges, and colliding safely with humans.
To achieve specific types of motion, roboticists have modeled their machines after crabs, jellyfish, termites, cockroaches and inchworms. Yet, with the exception of the jellyfish, these biomimetic robots are only capable of navigating in very proscribed environments. And because they’re hard and often heavy, the robots cannot generally share their workspace with humans.
MIT’s Computer Science and Artificial Intelligence Laboratory has made a significant step in achieving both safety and agility in the same device with a soft robot fish, made of soft silicone, that can perform sophisticated, agile movements and is safe for operation near humans.
“Because of their bodies’ capability to bend and twist, these robots are capable of very compliant motion. They’re capable of very rapid, agile maneuvers, which pushes the envelope on what machines can do today,” said Daniela Rus, director of the lab.
“They’re also inherently safe to be around,” she said.
Like a natural fish, the robot carries its computing power — including sensors and a wireless connection to a central computer — in a rigid head. The rest of the fish’s body is bendable.
Pioneers in soft robots have struggled to design machines that can hold their componentry on board without running into the same problems of rigidity and weight that limit conventional robotics. (Research is under way elsewhere to make electrical componentry flexible, obviating the need for a hard case.)
It’s plain old carbon dioxide that creates the fish’s movement. Within the fish head are two carbon dioxide canisters, and inside each side of the tail is a long, undulating channel that inflates with carbon dioxide. The canisters in the head each control one side of the tail by opening to different diameters for different durations to produce specific movements.
Like a real fish, the robot can perform a 100-degree escape turn in about one-tenth of a second.
Rus and her colleagues are pitching the fish as a coup for the emerging discipline of soft robotics.
“The maneuver is so fast and it’s got such high body curvature that it shows that soft robots might be more capable in some tasks,” said the fish’s designer, Andrew Marchese, an MIT electrical engineering graduate student.
Barry Trimmer, a biologist at Tufts University who edits the journal Soft Robotics, argues that “if we learn how to incorporate all these other sorts of materials whose response you can’t predict exactly, if we can learn to engineer that to deal with the uncertainty and still be able to control the machines, then we’re going to have much better machines.”
For instance, soft robots could potentially free up the machines to pursue more efficient forms of motion by eliminating the rigid constraint that they never risk colliding with humans, Rus said.
Inefficiency can dramatically slow the speed at which robots accomplish their tasks.
As with any proof-of-concept display, there’s room for improvement for the soft fish. After about 25 escape maneuvers, it runs out of gas. Marchese plans a second version that uses pumped water instead of carbon dioxide to inflate the channels in the tail. That model would be able to swim for about 30 minutes.
Images courtesy MIT