Sound sensors can identify the direction of sound and filter out noise

A team from Duke University, Durham, North Carolina, successfully developed a sensor prototype that could determine the direction of sound and extract sound from noise. Thereby in the future smartphones will be able to clearly distinguish between voice and environmental sounds and virtual assistants like Siri, Cortana or Google Now will be able to operate more efficiently.

Super sound sensors can identify the direction of sound and filter out noise

To create this type of sensor, the team used a material known as metamaterial - a metamaterial with properties not found in nature, more structurally dependent on structural components. creating and a signal processing technique named compressive sensing . The circular disc image sensor is made of plastic without the electronic or moving components. Instead, it is designed with honeycomb and thin leaf concentric structures, each with a distinctive loophole shape with different depths. These holes distort the sound waves and give the sensor special abilities.

"These holes act like soda bottles when we blow through the bottle. The amount of soda left in the bottle or the depth of the hole," said Steve Cummer , professor of electrical and computer engineering at Duke University . The sensor gap affects the high altitude they produce and this changes the sound in a vague but recognizable way. "

Distorted sounds carry many characteristics related to the leaves that carry the honeycomb structure. When the final sound is picked up by the microphone and processed by a computer, characteristic noises can be distinguished between noise. Therefore, the team thinks that sensors can split multiple tones at the same time from many directions due to distortion.

In the experiment, 3 different sounds were transmitted from 3 directions to the sensor prototype. The results show that the sensor can distinguish between these sounds and the ambient noise at 96.7% accuracy. The team hopes to be able to reduce the size of the sensor, which currently has a width of about 15 cm to be integrated into many devices. Besides applications on consumer electronics, sensors can also be exploited in many other areas.

Abel Xie - research leader said that sensors can be combined with medical imaging devices using sound waves (such as ultrasound waves) to increase screening efficiency . "With more information, it can improve sound fidelity and increase functionality for many applications such as hearing aids and cochlear implants One of the obvious challenges now is to shrink the device size. is a big challenge but not impossible, we are still looking for ways to achieve this goal. "