Researchers used bat ear design to develop a new approach to sound localization
Inspired by the bat's ear, Rolf Mueller, a professor of mechanical engineering at Virginia Tech, has created a bio-inspired technology that determines the location of a sound source.
The professor's development is based on a simpler and more accurate model of sound localization than previous approaches, which were traditionally modeled on the human ear. His work marks the first new idea for sound localization in 50 years. The research results are published in Nature Machine Intelligence.
“I have long admired bats for their remarkable ability to navigate complex natural environments using ultrasound, and I suspected that the unusual mobility of the animal's ears might have something to do with it”
A new model of sound localization
Bats navigate in flight using echolocation, determining how close an object is by continuously emitting sounds and listening for echoes. Ultrasonic signals come out of a bat's mouth or nose, reflect off elements of the environment, and return as echoes. They also pick up information from ambient sounds.
Comparing sounds to determine their origin is called the Doppler effect. In the human ear, this effect works differently. A discovery made in 1907 showed that people can locate sounds thanks to having two ears, receivers that send sound data to the brain for processing. Working with two or more receivers makes it possible to determine the direction of sounds containing only one frequency, and this will be familiar to anyone who has heard a car horn sounding nearby.
A horn is one frequency, and the ears work together with the brain to build a map of where the car is going. A discovery made in 1967 showed that when the number of receivers is reduced to one, a single human ear can locate sounds if different frequencies are present. In the case of a passing car, this could be a car horn combined with the roar of the engine.
According to Mueller, the workings of the human ear inspired past approaches to precise sound localization, which used pressure receivers such as microphones, combined with the ability to either collect several frequencies or use several receivers. Based on research with bats, the professor knew that their ears are far more versatile sound receivers than the human ear. This prompted his team to pursue the goal of one frequency and one receiver instead of multiple receivers or frequencies.
Creating the ear
Working from a single-frequency, single-receiver model, Mueller's team sought to replicate the bat's ability to move its ears. They created a soft synthetic ear, inspired by the horseshoe and Old World bats, and attached it to a string and a simple motor so that the ear fluttered in sync with the incoming sound.
Bats have ears that allow them to comprehensively transform sound waves, so nature's ready-made design was a logical choice. This transformation begins with the shape of the outer ear, called the pinna, which uses the ear's movement when receiving sounds to create many receiving shapes that direct sounds into the ear canal.
The biggest challenge the researchers faced with their single-frequency, single-receiver model was interpreting incoming signals. How do you turn incoming sound waves into readable and interpretable data?
The scientists placed the ear above a microphone, creating a mechanism similar to that of a bat. Rapid movements of the fluttering pinna created Doppler shift signatures that were clearly linked to the direction of the source, but they were difficult to interpret because of the complexity of the patterns. To address this, Mueller's team employed a deep neural network, a machine learning approach that mimics the multilayered processing found in the brain.
They implemented such a network on a computer and trained it to determine the direction of the source associated with each received echo signal. To test the performance of the system consisting of the ear and machine learning, they mounted the ear on a rotating platform that also included a laser pointer. Sounds were then played from a loudspeaker positioned in different directions relative to the ear.
Once the direction of the sound was determined, the control computer rotated the platform so that the laser pointer hit a target attached to the loudspeaker, locating it with an accuracy of half a degree. Human hearing typically locates sounds within 9 degrees when using two ears, and the best technologies can locate within 7.5 degrees.
“The capabilities go well beyond what technology currently offers, and yet all of this is achieved with far less effort. Our hope is to enable reliable and functional autonomy in complex outdoor environments, including precision agriculture and forestry, environmental monitoring such as biodiversity tracking, and defense and security applications”
There is still a lot of work to do before this becomes practical, but scientists are already seeing how their bat-based technology helps robots move through "complex" open spaces without much effort. This is another reminder that effective biomimicry is not necessarily based on our own biology.
