Google researchers have developed a system that can transform true wireless stereo (TWS) active noise canceling (ANC) headphones into heart rate monitors through a software update.
While smartwatches traditionally employ light pulses to monitor a user's heart rate, Google's research suggests using ultrasound emitted by an earphone's speaker driver to detect minuscule alterations in the ear canal's surface.
Google Introduces APG
In a paper titled "APG: Audioplethysmography for Cardiac Monitoring in Hearables," presented at MobiCom 2023, Google introduced a novel health sensing modality.
This method, known as Audioplethysmography (APG), allows ANC hearables to track physiological signals like heart rate and heart rate variability without the need for additional sensors or compromising battery life.
Google claimed that APG demonstrates resilience to motion artifacts, complies with safety regulations with a margin of 80 dB below the limit, remains unaffected by seal conditions, and is inclusive of all skin tones.
The deep ear artery, scientifically referred to as the arteria auricularis profunda, is responsible for supplying blood to the auditory canal. This artery intricately networks with smaller vessels that extensively permeate the canal.
Even subtle variations in the shape of these blood vessels, caused by the heartbeat and blood pressure, can lead to minute alterations in the volume and pressure of the ear canals. This unique physiological response makes the ear canal an ideal location for health sensing, according to the Google research team.
While prior research has delved into leveraging hearables for health monitoring, typically by integrating an array of sensors like photoplethysmogram (PPG) and electrocardiogram (ECG) with a microcontroller, this approach presents challenges.
It introduces additional costs, weight, power consumption, and complexity in acoustic design and poses form factor constraints to hearables. These factors collectively act as significant hurdles to their widespread adoption.
Google noted that Current Active Noise Cancellation (ANC) hearables employ a combination of feedback and feedforward microphones. These microphones have the capability to detect and record various biological signals both inside and outside the ear canal, opening up avenues for health monitoring.
However, it is worth noting that consumer-grade ANC headphones are equipped with high-pass filters. These serve to mitigate issues stemming from body movements or strong wind noise. But they also present a challenge when trying to incorporate health features reliant on passive listening to low-frequency signals.
APG circumvents these hardware constraints by emitting a low-intensity ultrasound probing signal through an ANC headphone's speakers. This signal triggers echoes, which are received via on-board feedback microphones. The subtle displacements of the ear canal skin and heartbeat vibrations modulate these echoes.
Read Also: Google Introduces AR Beauty Ads: Transforming Shopping Experience
Multi-Tone Approach
APG offers a unique advantage in transmitting multiple frequencies simultaneously, allowing for a higher-quality signal and greater resilience to motion. This multi-tone approach enables the development of advanced signal processing techniques.
Google conducted extensive user experience studies with 153 participants, demonstrating that APG consistently delivers accurate heart rate and heart rate variability measurements.
Unlike traditional PPG, which can exhibit variable performance across different skin tones, Google noted that APG proves to be resilient to variations in skin tone, sub-optimal seal conditions, and ear canal size.
"APG transforms any TWS ANC headphones into smart sensing headphones with a simple software upgrade, and works robustly across various user activities," Google said in a blog post.
"More importantly, APG represents new knowledge in biomedical and mobile research and unlocks new possibilities for low-cost health sensing," it added.
Related Article: Google Explores AI-Powered Search for Advertising Potential
