A New Bat-Like Sensor Captures Images Using EcholocationJune 10, 2020 by Antonio Anzaldua Jr.
What does IMERAI’s new sensor have in common with bats, dolphins, and whales? Echolocation.
Over the years, researchers have forged new electronic innovations by looking to nature—specifically in attempts to mimic animal capabilities. Most recently, we've seen this in a jellyfish-inspired, self-healing electronic gel for circuits and a wearable fabric sensor taking a note from octopus suckers.
But now, one company has turned its sights to bats, dolphins, and whales and their ability to orient themselves using echolocation. Scottish startup IMERAI has combined an ultrasonic echolocation sensor with artificial intelligence (AI) to determine surroundings. However, this sensor doesn’t use cameras or light to receive images; instead, it relies on sound.
IMERAI doesn't capture users' faces, so it can protect privacy. Image used courtesy of IMERAI
Founder and CEO of IMERAI Alex Bowen began his startup journey while attending Edinburgh Business Incubator at Heriot-Watt University. It was there that he developed a non-camera-based sensor that could add a visual element to voice-activated smart home assistants—without invading the privacy of its users.
How Does the Bat-Like Sensor Work?
The device works by bouncing ultrasonic waves off objects and detecting their relative distance and position. AI then processes that data and creates a detailed image of an environment.
“As with many problems, nature had a solution," Bowen explains. "In the wild, bats send out a screech and they listen for the echoes to understand distances and the location of physical objects. . . . Our sensors work in a similar way using echolocation to create a picture without any identifying data so that privacy is protected."
Founder and CEO of IMERAI Alex Bowen explains echolocation technology. Image used courtesy of IMERAI
IMERAI’s device utilizes a micro-electro-mechanical system (MEMS) microphone to recreate echolocation and gather data, which it then feeds to an AI system. It does this without collecting intrusive data (like individual faces) from cameras. By avoiding cameras and human oversight, this technology helps to protect user privacy—an ever-present concern with AI technology.
A Solution for Social Distancing
While MEMS microphones are already used in mobile devices, autonomous devices, and smart home applications, IMERAI feels its technology may provide the foundations for more sophisticated AI products.
One immediate application for this device may be in reopening workplaces, which may strive to uphold social distancing measures. Bowen says his company's bat-like sensor can count how many people are in an office and determine how far apart each individual is from others.
Waveform captured from IMERAI sensor. Image used courtesy of IMERAI
Another long-time application for this technology may crop up in assisted living centers. Bowen explains, “This could be game-changing for dementia sufferers and others with assisted living needs, allowing their movements to be monitored and any deterioration to be picked up more quickly.”
IMERAI can be useful in these scenarios because it doesn't capture sensitive data—that is, images of people's faces. Rather, it generates anonymous videos without identifying users.
Echolocation vs. LiDAR
How will IMERAI’s sensor measure up to current camera-based AI technology?
LiDAR sensors, which we most often associate with autonomous vehicles, capture thousands of images by illuminating targets and measuring the return time for light to bounce off an object. The wavelengths are analyzed to make digital representations of the selected targets. This technology has been modified for various applications—for instance, in archeological surveyance—by increasing the number of light pulses or shortening the pulse for finer resolution.
IMERAI's ultrasonic echolocation sensor, on the other hand, measures wavelengths of sound to establish a 3D image and recreate its surroundings. "Unlike alternative 3D imaging methods, our device is completely static and has no moving parts," IMERAI adds in its incubator description.
IMERAI's sensor, which is said to securely ring-fence computer vision in any product. Image used courtesy of IMERAI
IMERAI says its technology can complement cameras, which can drain power in always-on mode. It can also vary its data output, acting as simply as a distance sensor for low-power applications or providing full 3D video.
IMERAI’s open-air acoustic imaging technology can also replace cameras in robotics and internet-connected devices.
Immersive AI With Echolocation
The company claims that its echolocation sensor marks a new chapter in depth-perception technology because it enables computer vision without optical components.
Engineers working with neural networks may, in the future, input echolocation data to help their systems recognize patterns, capture routines, and adapt to change.
While IMERAI is still in its infancy as a company, it has recently attracted enough investors to build out a full engineering team to continue developing this technology.
Have you ever looked to phenomenon in nature as inspiration in circuit design? How? Share your experiences in the comments below.