This is according to Audio Analytic, Vesper and Ambiq Micro, who have established a partnership to deliver the AI sense of hearing to battery-powered consumer electronics.
Example applications include the ability to detect and react to smoke alarms, CO alarms and windows being broken to boost smart home security and consumer peace-of-mind, as well as many more opportunities outside the home, says the team.
Tests were conducted by Audio Analytic using a combination of Vesper’s piezoelectric MEMS microphone and Ambiq Micro’s subthreshold power-optimised technology (SPOT) processors. During tests, using real-world audio data from Audio Analytic’s Alexandria dataset, the combination of CPU, microphone and embedded, lightweight AI software could run for up to five years on a number of battery configurations.
Battery-powered devices already successfully launched in the market include portable smart home cameras, smart locks and stick-up motion sensors. Significant innovation means that these devices can deliver up to one or two years before the batteries need to be replaced. These products, from market leaders Blink, Nest, Ring and Yale, have set consumer expectations around battery life. As a result, any new functionality can’t impact on these figures. By upgrading their hardware, consumer tech brands could enhance functionality and value through sound recognition without impacting on battery performance, says the team behind this latest development.
For example:
- A stick-up battery powered security camera could combine the power of sight and sound and still last for two years on two lithium AA batteries
- A smart lock, running on four standard AA batteries, could enhance its security and safety applications and still last for at least a year
- A wall-mounted motion detector could hear the presence of suspicious activity even if it could not see it for at least two years on a tiny CR123 battery
Due to its extensive work and research in sounds, Audio Analytic says it was able to use real-world data from a diverse range of home environments in order to accurately model battery performance.
“We’ve built the world’s largest audio dataset for machine learning called Alexandria. As well as target sounds, the dataset also contains environmental sound data from volunteers. This means that we can accurately model these environments and help our customers meet their targets for battery life,” comments Audio Analytic CEO and founder Chris Mitchell.
Beyond the growing smart home market, battery-powered sound recognition opens up opportunities for other applications, including headphones and large-scale IoT sensors. For example, a battery-powered sense of hearing enables headphone manufacturers, who are targeting around five hours of battery life, to add innovative context-aware sound recognition to products, thus enabling wearers to get lost in their music but still stay in touch with the world around them.
Mitchell adds: “Our research shows that consumers are largely positive about AI’s role in helping them. They also want flexibility and an easy route to installation. Devices that require wires or even professional installation can present barriers to adoption. So if you can offer cutting-edge AI and still enable consumers to immediately benefit from their purchase then we see that as a win-win.
“Companies like Amazon-owned Blink and Ring, Nest and Yale are all selling battery powered devices, so we expect to see this trend increase. Adding AI capabilities like sound recognition enables these devices to evolve from smart to intelligent.”
“Extended battery life in connected devices is the future of IoT,” Matt Crowley, CEO of Vesper, adds. “Our piezoelectric MEMS microphones, combined with AI and power-optimised technology, will breed a new generation of intelligent devices. Our proprietary Zero Power Listening enables devices to conserve battery by remaining in full power down mode until awoken by a wake word or another acoustic event, increasing battery life from hours to years.”
“Meeting steep power consumption demands remains the largest barrier to truly enabling AI on battery-powered edge devices,” says Aaron Grassian, VP of Marketing at Ambiq Micro. “Audio Analytic’s unique sound recognition technology running on Ambiq Micro’s ultra-low power SPOT-based processors removes this barrier and proves that energy-efficient hardware platforms with enhanced user experiences can be built without sacrificing product quality, size, and battery life.”
