Team Investigates Why Drones Can’t Handle Turbulence Like Birds
Birds have little problem flying smoothly and maintaining a fairly direct path despite turbulent air such as found in urban “canyons,” but even the best drones can’t do that. A team at Stanford University is working to understand how birds dynamically adjust their wing and body configuration to do this, by using a custom wind tunnel and a projected image-capture system (see figure).
In the system built by assistant professor of mechanical engineering David Lentink and his students, the bird thinks it’s flying forward, but is really stationary as the air in the tunnel moves, thus allowing a camera to more easily capture the bird’s flight. The bird is effectively “stuck” in the tunnel, and not only experiences basic laminar flow, but also fully controllable turbulence generated by a set of moving vanes.
“We’re trying to figure out how birds are capable of flying so well in these complex, turbulent environments, and a lot of that comes from how they deform the shape of their wings, left versus right, to adjust to gusts quickly,” says Prof. Lentink.
There’s another advantage to the tunnel arrangement. The usual bird-tracking approach requires attaching markers to the bird’s extremities for the image capture and subsequent analysis, which distorts the bird’s shape, motion, and airflow. Also, that approach can’t fully reconstruct an entire wing surface at high resolution. Another standard method uses projection of patterned light onto the bird, which enables automated analysis. That’s too slow to record most actual bird flight, though.
By using a wind tunnel in which the bird of interest always remains in view despite air flow and turbulence, researchers at Stanford University can employ an advanced imaging system to capture and study how the bird morphs and adjusts itself to maintain smooth flight despite these conditions. (Source: Stanford University)
Getting it in Real Time
Here, however, the bird is stationary with respect to the wind-tunnel viewing window. Therefore, both the image and dynamic configuration can be captured in real time. A video camera is synchronized with a projection of two overlapping light patterns. One is a dense grid that covers the surface of the bird, and provides a high-resolution image; the second is a set of unequally spaced lines to ensure that no two areas of the field look alike. The body of the bird functions like a projector screen and the straight projected onto the bird’s body deform based on its shape.
This isn’t the first structured-light system built by the lab, but this latest version automatically resolves body-shape changes at high speed, and does so in high resolution. “The great thing about this system is it’s the first fully automated, high-speed reconstruction of birds in the world,” says Marc Deetjen, a graduate student in the lab, and the lead author on a paper published in the March 27 issue Journal of Experimental Biology.
The test “subject” used was a four-year-old parrotlet from whom the team gained remarkable insight. They computed the bird’s effective aerodynamic angle of attack and found it was consistently between 55° and 75°in the first downstroke, and between 45° and 60° in the second. This is contradictory to guidelines for conventional aircraft—most stall when the angle of attack reaches about 15° as the associated drag separates the airflow from the wing, resulting in greatly reduced lift. Here, the researchers concluded that the bird is actually supporting its body weight using drag, which is oriented upward. The lift it generates is rotated forward so that it functions as a thrust vector.
