The amphibious aircraft developed by EPFL and MIT is inspired by diving animals: it takes off, dives, resurfaces and returns to the air without changing propulsion

Crossing the boundary between air and water represents one of the most complex challenges in robotics. The two environments have profoundly different physical properties: water is about a thousand times denser than air and requires completely different propulsion, control, and energy consumption strategies. For this reason, amphibious robots developed to date have generally used separate systems for flight and swimming, or have had to rely on propellers, floats, or launch pads to return to the air.

A group of researchers from the Federal Polytechnic School of Lausanne (EPFL) and Massachusetts Institute of Technology (MIT) has instead demonstrated that it is possible to address both modes of locomotion with a single mechanical architecture. The result is the Flapping-wing Aerial-Aquatic Vehicle (FAAV), a bioinspired robot that can swim underwater, break the surface, and immediately resume flight simply by continuing to flap its wings. The study was published in the scientific journal Science.

The project stems from the observation of a small group of seabirds that can navigate efficiently in both environments. Loons, gulls, puffins, and petrels use the same wings to generate lift during flight and thrust during swimming, adapting their movement to the characteristics of the fluid they are navigating. Translating this capability to a robotic system required years of study of animal biomechanics and numerous experimental trials.

The main research manager is Raphael Zufferey, now an assistant professor at MIT after a research period at EPFL. The initial goal was not so much to build a new amphibious drone as to understand the principles that allow diving birds to overcome a transition that, from an engineering perspective, appears extremely difficult.

Flexible wings and a single propulsion for two opposite environments

The FAAV weighs less than 300 grams and consists of a waterproof central fuselage, two flexible wings and a swivelling tail. Inside the body there is a battery and a waterproof electric motor which, through a crankshaft, gives the wings the oscillating motion necessary for both swimming and flight. The wing membranes are also covered with hydrophobic nanoparticles, which favor the rapid detachment of the water during emergence.

One of the most interesting aspects concerns the modularity of the system. The researchers have created three different wing configurations, with openings of 60, 80 and 100 centimeters, to understand which combination produced the best compromise between lift, thrust and maneuverability. The tests showed that the intermediate solution 80 cm ensures the most balanced behavior both in water and in the air.

Wing flexibility has proven equally crucial. Overly rigid structures improve flight but increase hydrodynamic drag; excessively deformable wings facilitate swimming but don't produce sufficient lift after surfacing. Optimal behavior thus results from a compromise very similar to that evolved in seabirds over millions of years.

Tests between the university's experimental tank and Lake Geneva

To verify the functioning of the robot, the group carried out a long series of tests first in an experimental tank and then in Lake Geneva, in Switzerland. During each test, the vehicle was positioned about half a meter deep and programmed with different wing beat frequencies and different tail inclinations. The goal was to identify the combinations capable of ensuring a stable transition from water to air without loss of control.

The results showed that the robot can swim at speeds close to one meter per second with a frequency of approximately five beats per second, while in flight it reaches approximately six meters per second maintaining a similar frequency. The performance is surprisingly close to the values ​​observed in diving birds of comparable size.

The most delicate phase is crossing the surface. To prevent the tips of the wings from hitting the water during the flap, the robot must assume an inclination of approximately 70%A lower angle prevents takeoff, while a higher angle increases the risk of the aircraft immediately falling back into the water. The combination of wingspan, oscillation frequency, and tail pitch therefore creates an extremely precise balance that allows the FAAV to transition from one environment to another without interrupting the movement of its wings.

Take-off is successful without the aid of legs or separate thrusters.

The combination identified by the researchers allows the FAAV to complete the entire sequence without the use of additional propellers, floating floats, or other thrusters. After swimming beneath the surface, the vehicle orients its body upward using its powered tail, continues to beat its wings, and uses the accumulated thrust to surface. As the fuselage rises above the water, the same movement generates the lift needed to continue in the air.

The ability to complete this transition without using its legs also distinguishes the robot from the behavior of many diving birds. Several species, when they need to regain altitude, paddle on the surface to increase their speed before taking off. The prototype developed between Lausanne e Cambridge Instead, it manages to directly transform the underwater heartbeat into a take-off maneuver.

Raphael Zufferey He explained that the very absence of an apparatus equivalent to legs was one of the central questions of the research.

"If you look at birds, most of them have to paddle on the surface with their feet to take off. The question was whether it was necessary to do the same with robots, but we discovered that it isn't. No one had yet succeeded in flying out of the water using their wings."

The result does not depend on a single component, but on the interaction between wing geometry, membrane elasticity, oscillation frequency, and tail trim. A small variation in any of these parameters can prevent surfacing or make the transition to flight unstable. The FAAV therefore functions as an integrated system, in which the mechanics must simultaneously adapt to hydrodynamic resistance and aerodynamic requirements.

The comparison with animals also allowed for the controlled study of some principles that are difficult to isolate in nature. Birds continuously modify their posture, joints, and heartbeat amplitude; the robot, on the contrary,

It allows you to set precise values ​​and observe how each choice affects speed, stability and the ability to cross the surface. The prototype thus becomes not only a possible operational means, but also a tool for analyzing the biomechanics of amphibian locomotion.

From icebergs to ports, the missions already imagined at sea

The applications considered by the group mainly concern theocean exploration and data collection in areas difficult or risky for traditional vessels. A lightweight vehicle could launch from the coast or a ship, quickly fly to the area to be surveyed, dive to take a measurement or collect a sample, and then return to base carrying the collected information.

Examples include: iceberg, port infrastructure, and areas frequented by cetaceans. In each of these environments, the ability to alternate between air and water would avoid having to use two different means, one airborne and one underwater, or manually retrieve the device after each dive.

Zufferey described a possible mission in which oceanographers, marine biologists, and coastal communities would launch the robot from a boat or directly from shore.

"Our vision is for the robot to fly near an area of ​​interest—for example, an iceberg, a port facility, or a pod of whales—dive to take a measurement or collect a sample, and then return to deliver the data at a fraction of the cost of traditional methods. It could then depart for another dive."

The reference to costs does not yet equate to a complete economic evaluation. The FAAV remains a experimental prototype The research does not provide an operational comparison with oceanographic vessels, autonomous underwater vehicles, or commercial drones. The potential advantage stems primarily from the reduction in launch and recovery operations and the ability to cover distances in flight that a small underwater vehicle would travel more slowly.

Even weighing less than 300 grams represents both an opportunity and a limitation. Such a lightweight device can be easily transported and deployed, but it has limited capacity for batteries, measuring instruments, and sample containers. Scientific use will therefore require sensors compatible with the available weight and effective protection from salt water, collisions, and wave stresses.

Tacking, wind and rough waters are the next test benches

The team is now modifying the wings to control turns, in addition to the vertical movement already tested. The current configuration demonstrates the continuity between swimming, surfacing, and flight, but a vehicle intended for real-world missions will need to orient itself precisely, correct its course, and react to disturbances without depending on calm conditions.

The subsequent tests will in fact concern situations closer to those encountered in the open sea: choppy waters, turbulence during emergence, and wind during the aerial phase. Tests in Lake Geneva have verified the operating principle, but irregular waves and gusts can alter the angle at which the robot reaches the surface, rewetting the wings, or destabilizing the transition from one fluid to another.

It also remains to be determined how long the vehicle can operate, what scientific payload it can carry, and how reliably it can repeat numerous consecutive transitions. These elements separate the experimental demonstration from an actual oceanographic application and will need to be measured in more extensive campaigns.

The study involves, in addition to EPFL and MIT, researchers from Northwest Indian College in the United States. The work was supported in part by a grant from the Marie Sklodowska-Curie Actions and from Swiss National Science Foundation. The publication on Science It documents above all a mechanical achievement: a single wing apparatus can produce the necessary thrust underwater and the required lift in the air, without interrupting movement during surfacing.

Before this architecture can become a standard tool for oceanographers and marine biologists, more comprehensive guidance systems, greater resistance to environmental conditions, and a load capacity compatible with measurements and sampling will be needed. However, the FAAV has clarified the starting principle: to move between sea and sky, it's not essential to combine two different vehicles, as long as the wings, tail, and attitude control are designed to handle both environments.

Here are three insights that might interest you:

A brand new drone for ultra-fast transitions between air and water
Award to ProteusDrone, the soft robot with mutant shapes
Zebrafish-inspired robot unravels behavioral circuitry