Drones inspired by birds of prey could navigate hard-to-access locations

Engineers at the University of Surrey are developing fixed-wing, unmanned aerial vehicles (UAVs) capable of performing agile manoeuvres, such as perching or obstacle avoidance, by studying the flight behaviour of owls and other precision flyers.

The project, called ‘Learning2Fly’, aims to overcome key limitations of conventional drones, particularly in environments where space is tight and wind conditions are unpredictable. 

Unlike standard rotary-wing drones, which are highly manoeuvrable but energy-intensive, fixed-wing drones are far more energy-efficient and capable of covering
longer distances, making them ideal for applications, such as wind turbine inspections at sea.

While they have typically lacked the agility needed to fly safely and precisely through turbulent or cluttered airspace, the Surrey team’s project could allow this new class of UAVs to operate with far greater control and adaptability by harnessing wing aerodynamics. 

“Nature has already solved many of the challenges we face in drone flight,” Dr Olaf Marxen, Senior Lecturer.

“Birds of prey can perform incredibly precise
manoeuvres in complex environments, and we’re using those lessons to make fixed-wing drones smarter, more agile and better suited to cities with tall buildings or rapidly changing wind conditions.

“We’re combining experimental flight data with machine learning to help drones predict and control their motion in real time to imitate a bird’s typical flight path.

“Traditional simulations such as computational fluid dynamics fall short in turbulent environments and are prohibitively expensive, so our next step is refining the predictive model
and testing outdoors, bringing us closer to deployment.”

Instead of relying on complex computer simulations, researchers are testing the manoeuvres in real-world experiments using Surrey’s motion capture lab.

A number of lightweight prototypes have already been built and tested, some of which were adapted from commercial toy planes, to track their motion in 3D using onboard sensors and high-speed cameras.

The data collected is being fed into a machine learning model, helping the team predict drone behaviour without relying on
conventional aerodynamic simulations. 

“We’ve already presented some of our early findings, and it’s exciting to see how well the drone performs even at this stage,” said Owen Wastell, University of Surrey PhD student and project co-lead.

“It’s humbling that in an era of advanced machines and technology, we’re still looking to the natural world – and one of the oldest living species on the planet – for inspiration.”

With further testing planned for outdoor environments, researchers hope the project will lay the groundwork for a new generation of agile, energy-efficient drones guided by nature.