Adaptive listening sensor ‘learns’ background noise, battery life rises - Embedded.com

Adaptive listening sensor ‘learns’ background noise, battery life rises

Piezoelectric MEMS technology company Vesper has announced its next generation microphone based on a listening sensor which ‘learns’ from background noise so that it stays in hibernation longer and hence provides a 10x battery life extension.

Targeting the low power always-listening device market, Vesper said the new VM3011 MEMS microphone is the first intelligent digital microphone using its adaptive ZeroPower Listening (ZPL) technology. Adaptive ZPL automatically learns the acoustic characteristics of the environment in real-time, allowing the system to ignore background noise and only wake in response to keywords or other desired acoustic triggers. This allows the system to hibernate for up to 90% of the time and consequently extend battery life.

Vesper’s CEO Matt Crowley explained, “Existing always-listening systems run continuously and must be plugged into an electrical outlet. The VM3011 creates a new paradigm where these systems can now run for months or years on battery power. This technology will power a new class of products such as always-listening far field TV remotes, always-listening smartwatches, home security cameras and doorbells with extended battery life and many other systems. Vesper has already engaged with lead customers who plan to roll out products with the VM3011 later this year.”

Target applications for the VM3011 include internet of things (IoT) products, portable smart speakers, smartwatches, far field remote controls, security cameras, glassbreak sensors, automotive, medical devices and other battery powered systems.

In the VM3011, the microphone constantly listens to the environment, consuming 10 µA of current. In adaptive mode, it is continuously measuring the average sound level of the environment and adjusts the internal wake on sound threshold comparator until it converges to the threshold margin configured using the I2C register. This margin can be set to a range 6-18 dB above the average acoustic noise level. When the required threshold is reached, the microphone sends an interrupt signal to the external codec or voice processor using a digital output pin (DOUT).

Vesper VM3011 adaptive mode
Adaptive mode continuously measures the average sound level of the environment and adjusts the internal wake on sound threshold comparator until it converges to the threshold margin. This margin can be set to a range 6-18 dB above the average acoustic noise level. When the required threshold is reached, the microphone sends an interrupt signal to the external codec or voice processor using a digital output pin (DOUT). (Image: Vesper)

The key benefit of tracking ambient noise is the extension of hibernation time. Vesper said that an always listening remote can get 18 months battery life using the new device.

In addition to this adaptive listening mode, the VM3011 also features a fast response time of around 200 µs, requires no mesh and features IP57 rating. The response time compares against around a typical 10-20 ms for capacitive sensors; this, according to Vesper means that the devices wakes within microseconds of the sound event to capture the full wake word, improving wake word detection accuracy of batter-operated devices.

The elimination of the need for an external protective mesh translates to saving device manufacturers’ bill-of-materials (BoM) costs, as well as increasing the acoustic performance of the end application; Vesper said a mesh can lead to a -20db or 90% signal loss.  In dust and water tests Vesper claims it was the only microphone company to achieve a zero-failure rate after being fully submerged in water and subject to extensive dust/harsh environmental testing.

Crowley said that Vesper already has some reference designs and expects mass production starting in September 2020. Early customers are in remote controls, wearables (smartwatch) and glass break detection. He added its first generation VM1010 has already been shipping with three large customers and a dozen or so smaller ones. Many of these are in far-field remotes and home security, and when asked about volumes shipped, he said it was greater than 10 million and less than 100 million.

The VM3011 delivers high acoustic performances through a bottom port digital MEMS microphone with an industry standard package size of 3.76 x 2.95 x 1.3 mm. The microphone integrates both the ASIC and the piezoelectric MEMS transducer in the same package. A digital PDM interface is available to stream the audio data while an I2C interface allows “on the fly” configuration of the adaptive ZPL.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.