News & Analysis
/
Article

Neural network offers insights into aerodynamics of flapping insect flight

MAR 21, 2025
Exploring variations in both wing kinematics and wing geometries produces optimized values that closely match flight behavior of fruit flies.
Neural network offers insights into aerodynamics of flapping insect flight internal name

Neural network offers insights into aerodynamics of flapping insect flight lead image

Flying is a complex endeavor. Humans have invented several wildly different ways of getting ourselves airborne, and in animals, flight has independently evolved at least four times. Among those, insects have a particularly complicated set of strategies that offer them unparalleled mobility.

To better understand the intricacies of insect flying dynamics, Lee et al. employed a neural network to predict the aerodynamics of hovering flapping flight over a wide range of wing shapes and motions.

“Our approach is, to our knowledge, the first one that works for both various wing kinematics and geometries, as previous studies varied one while keeping the other fixed,” said author Haecheon Choi. “This capability allows us to optimize the wing kinematics and geometry simultaneously, providing a more holistic way to improve the aerodynamic performance.”

This comprehensive spectrum of evaluated models allowed the authors to determine the optimal wing kinematics and geometry — including its planform, its shape as viewed from overhead — for sustained hovering flight.

“Interestingly, the optimal wing motion and planform shape identified by our model remarkably resemble those of a fruit fly,” said Choi. “Furthermore, the optimal solutions with and without elastic energy storage match closely with the upstroke and downstroke of the fruit fly, respectively, providing potential insights into its asymmetric stroke pattern.”

In addition to learning more about fruit fly flight, this research could help guide development of flapping wing micro-aerial vehicles. These insect-inspired robots can execute precise aerial maneuvers and could be used for search and rescue or environmental monitoring. The authors plan to explore more types of flapping flight with the goal of improving the efficiency of these miniature robots.

Source: “Neural-network-based optimization of the wing kinematics and geometry of a hovering flapping flight,” by Hyeoksu Lee, Sehyeong Oh, and Haecheon Choi, Physics of Fluids (2025). The article can be accessed at https://doi.org/10.1063/5.0256748 .

More Science
APS
/
Article
A drone-borne photon counter can provide a new view into environmental health.
AAS
/
Article
JWST observations of the young star cluster IC 348 revealed extremely low-mass brown dwarfs with signatures of hydrocarbons in their spectra: meet the members of the proposed “H” spectral class.
/
Article
Molecular dynamics model uses topological order parameter to show stratification of water molecules based on density.
/
Article
The proposed next generation of gravitational wave detectors will require hundreds of kilometers of ultrahigh vacuum beam pipes, a costly endeavor.