MIT’s 250-Gram Puffin Robot Swims Underwater, Then Flies

Imagine a seabird diving into a lake, chasing fish underwater, and returning to the sky without looking back. Engineers at MIT and EPFL recently created a robot that swims and flies just like that, and weighs less than a soda can. No propellers, no legs, no flipping tricks, it swims, dives and flies with two flapping wings.

A robo-bird at first glance

Nature understood this first

About 100 species of birds already divide their days between water and air, among them are loons, gulls, puffins, and petrels. They dive to chase fish, then drag them back into the sky seconds later. This robot copies that entire loop, and the team calls it the first bird machine to swim, dive, launch, and flap on its own. At about 250 grams (8.8 oz), it sits in the same category as seabirds.

The way it falls is from water to air

Water is 1,000 times denser than air, so many machines are tricked by handling different systems for each machine. This robot skips that and relies on dynamic wings and carefully tuned flapping. Under water the wings bend to shorten each stroke and protect the engine, then stiffen enough to hold the robot when it’s airborne.

The pace changes with the work. Moving in any direction, the robot flaps its wings five to six times per second. That keeps it swimming at about 1 meter per second and flying at about 6 meters per second. Leaving the water is different, because there it has to rise to about 10 beats per second to break the surface.

Inside the open body sits a battery and a waterproof motor that drives the crankshaft, pumping the wings up and down. Those wings are flexible nylon fabric reinforced with carbon fiber struts and coated to shed water, and the tail is motorized to tilt the robot up or down. The entire robot also moves in neutral, so it hangs in place underwater instead of drilling or sinking, which saves a little battery.

The hardest part is leaving the water

Swimming works. Flying works. The problem lies in the handoff between them, immediately the robot has to break and take off from the air by the power of the wing alone. Miss that beat and the rest doesn’t make sense, which is exactly why the previous bird-sized robots never managed it.

The team found that the robot pulled it off most reliably when the wings hit the middle of stiffness, curved enough to move through water but stiff enough to lift it into the air. The exit angle turned out to be the deciding factor. A pitch of about 70 degrees keeps the wingtips from sticking up when climbing, and anything steeper steers the robot back into the water.

What we teach about diving birds

Payment here is not just something that machines do. Since you can scale and maneuver the robot in ways you never could with a live puffin, it serves as a testing ground for open questions about diving birds. One example: the team hypothesizes that birds that thrust their wings underwater may be chasing speed and sharp maneuvering rather than saving effort. That search is almost impossible to detect in a wild animal, but it is easy to distinguish it from a machine.

Birds and robots also live in the same efficiency band, a metric called the Strouhal number that ranges from about 0.2 to 0.4. And where the bulkier diving birds tend to push off with their legs, this one leaves that step and still lifts off the surface.

If flying beats swimming

Data hides the limitation as well. Passing about 15.5 m (51 ft), flapping in the air costs the robot less than paddling the same area underwater. So on long enough trips, it pays to get out, sail up, and back down instead of going all the way in.

It is tested in the pool, and no feet are needed

The team built three sets of wings, then ran the robot first in a water tank and later out into Lake Geneva. Medium wings, about 80 centimeters wide, hit the sweet spot. The real surprise came at takeoff. It can come straight out of the water without paddling, a movement that many wading birds rely on for air.

If you look at birds, most birds need to paddle their feet up to take off. And the question was, do we need the same for robots? And it turns out we don’t.

That’s Zufferey again, and it’s a discovery that breaks new ground, because no one has come out of the water on their own wings before.

What’s next for the robo-bird

The system isn’t autonomous yet, and planes until now have operated with programmed flaps and tail angles rather than live control. Next, the team wants wings that can flip and flap, and be tested in extreme conditions such as turbulent water and wind. In addition, they see a low-cost marine science instrument, which reaches the most dangerous areas on conventional ships.

Our dream idea is for marine experts, marine biologists, and members of coastal communities to launch this robot from a boat, or from the shore, and it can fly near an area of ​​interest, such as an iceberg or a harbor area, or over a mass of whales. It can dive to take a measurement or collect a sample, and fly back to deliver data at a fraction of the cost of conventional methods.

That quote comes from Raphael Zufferey, an assistant professor of mechanical engineering at MIT, who worked on the project with Moritz Hüsser, and the research was published in the journal. Science.