Led by mechanical engineering professor Rolf Mueller, a team of researchers at Virginia Tech developed a new, bat-inspired technology to help robots accurately determine a sound’s origin. The team hopes that its new technology will improve robotics for agriculture, environmental surveillance, and of course, defense and security.
Scientists and engineers base most sound location technology on human hearing, which is relatively inaccurate. Humans rely on both ears to determine a sound’s origin with a 9 degree accuracy, while bats can pinpoint sound within half a degree using either one of their ears.
Both humans and bats determine a sound’s origin through the Doppler effect, a phenomena where a sound’s frequency (and therefore its pitch) increases or decreases as you approach or move away from a sound source (the Doppler effect doesn’t happen when you’re standing still, only when you or the sound source are actively moving). Because bat ears are constantly flicking and fluttering, they can “scan” a sound for its Dopplar shift signature and determine its location with more accuracy than a human.
The new sound location technology is, at its most basic level, a replica of the bat ear. Rolf Mueller and his team crafted a synthetic bat ear that moves and flutters, relaying a sound’s Dopplar signature to a small microphone. Then, a neural net specially trained to parse Dopplar shift signatures determines the sound’s origin with incredible accuracy.
As of now, the sound locating system developed by Rolf Mueller and the Virginia Tech team relies entirely on bat anatomy. Future improvements could eliminate the need for a synthetic bat body part, but there’s a good chance that we’ll see autonomous robots with wiggly, fluttering bat ears.
Source: Virginia Tech Daily via Engadget