Studying bats to improve sensor design


Echolocation, the process bats use to track their prey, is fairly well understood. While looking for targets, a bat emits a high-pitched sound; the sound travels through the air, bouncing off anything in its path. Meanwhile, the bat is listening for the echoes that bounce back, and can figure out the location of a moving target based on how quickly or slowly the echoes return. Using this technique, a bat can track and catch flying prey even in utter darkness.

Despite many years of researchers studying echolocation, however, one question remained unanswered: if bats track moving prey by gathering information from echoes, how do they track prey that isn’t moving, like a sleeping dragonfly sitting on a leaf? Without any movement on the part of the dragonfly, its echoes would simply “blend in” with those from its surroundings; a bat would have no way of knowing it was potential prey. It could simply be another leaf, a twig, or other inedible object. Still, bats have been observed locating silent and unmoving insects amid clutter and moving in for the kill.

Now, researchers may have figured out how. Starting with a high-speed video of a bat capturing a still dragonfly sitting on a leaf, Roman Kuc, professor of electrical engineering, and colleague-son Victor Kuc noted that the flapping of the bat’s wings caused the dragonfly’s wings to move as well, while the more massive leaf moved much less. Speculating that this induced movement, which occurs in sync with the movement of the bat’s wings, might contain enough echolocation clues to let the bat track the otherwise still prey, the researchers set out to model the induced movements and their effect on returning echoes.

Using robot sonar, a real dragonfly specimen, and plastic leaves with reflection properties similar to those of real leaves, the researchers measured the induced movements in response to air puffs generated by an airbrush to simulate the forces produced by a bat’s wing. They then studied the resulting echo waveforms. While echoes from the leaves fluctuated negligibly, those from the dragonfly wings exhibited dramatic changes. Differences that were in sync with the bat wing’s movements were apparent in the waveforms, and the researchers suggest that bats can use these differences to detect movement induced by their own wings, locating prey that is otherwise still and silent.

“The remarkable behaviors of biological sensing systems provide cues for improving the perception capabilities of robots through the design of biomimetic sensors,” says Kuc. “Understanding the fundamental sensing limitations then allows us to design hyper-biomimeticsensors that do not have the constraints that are imposed on biological systems.”