Fly Like an Albatross: Embry-Riddle Drone Research Aims to Mimic Birds

When eagles, hawks and albatrosses fly, they have the capability to catch a ride on winds and thermal currents, effectively harvesting energy from the atmosphere and converting it to flight.

Could drones be programmed to soar in the same way — not only to sense air flow but to read the weather and choose the most energy-efficient flight paths in real time?

Embry-Riddle researchers are aiming to accomplish just that while participating in the Defense Advanced Research Projects Agency (DARPA) Albatross program, said Michael Kinzel, associate professor in the Department of Aerospace Engineering.

“The Albatross project is about designing intelligent gliders that can fly longer and more efficiently by learning from nature — specifically mimicking how birds such as an albatross use wind to soar,” said Sweety Sarker, Ph.D. candidate in the Department of Aerospace Engineering. “The goal is to combine real-time weather forecasting with flight planning, so the glider can adapt its path to take advantage of the wind.”

The Embry-Riddle component of realizing these gains is weather prediction and is being developed by Kinzel, Sarker and Brendon Cavainolo, a postdoctoral scholar in Aerospace Engineering.

The effort “utilizes weather forecasts to build more detailed, microscale models that directly capture wind details around topography and waves,” said Cavainolo. These microscale models are being explored to either communicate with the glider from servers, or to “run onboard, where the models guide the glider to make decisions in real time,” Sarker said.

“Modern machine-learning algorithms combined with a range of computational fluid dynamics methods are how we are combining conventional and modern technology to help us succeed,” said Cavainolo.

Kyle Collins, research assistant professor in the Aerospace Engineering Department and director of the Eagle Flight Research Center (EFRC), is a co-principal investigator on the project and leads a team that will conduct developmental testing of the system. The EFRC team is currently evaluating sensing technology for the marine environment and plans to support flight test campaigns over water. EFRC’s flight research experience and access to such marine environments support the marine-flight mission goals of the program,” Collins said.

Embry-Riddle received $1.4 million from DARPA to conduct this research, which also includes the flight testing over water. The researchers say soaring technologies would be most beneficial with the winds and weather influences found over and near the ocean.

Sarker is developing physics-informed neural networks (PINNs) to model the complex aerodynamic flows surrounding the glider. In this application, these networks enable machine-learning predictions of airflow by embedding flight parameters governed by the laws of physics. “My work focuses on building models that capture airflow around the flight environment,” Sarker explained. “For example, PINNs can approximate local wind fields, which in turn support dynamic soaring maneuvers.”

According to Kinzel, the team expects to achieve predictions in near real-time, allowing for the exploration of a variety of methods. The research program is led by the Raspet Flight Research Laboratory at Mississippi State University and is also in collaboration with the University of Texas at El Paso.