Problem
Research questions and friction points this paper is trying to address.
heart rate sensing
motion artifact
photoplethysmography
wearable devices
free-living activities
Pixel Watch: Robust Heart Rate Sensing from Multipath PPG and On-Device Deep Learning Trained on 10,000 hours of Free-Living and Fitness Data
🤖 AI Summary
This work addresses the challenge of motion artifact interference in heart rate monitoring during intense exercise and free-living conditions using smartwatches. To this end, it presents the first integration of multi-path photoplethysmography (PPG) hardware with an on-device deep learning model on a Google smartwatch. Trained on 10,000 hours of real-world data, a 15-layer temporal dilated convolutional network—comprising approximately 300,000 parameters—processes ten optical signals in real time to produce a 1 Hz heart rate estimate. The proposed approach substantially surpasses the performance ceiling of conventional signal processing methods, achieving 95% limits of agreement of [−10.34, 8.66] BPM on a fitness validation set and [−6.57, 7.48] BPM on a free-living validation set, with significantly lower error than previous-generation devices.
📝 Abstract
The Pixel Watch 2 (PW2) is the first Google smartwatch to combine multipath photoplethysmography (PPG) with deep learning-based heart rate inference, designed to significantly improve sensing accuracy during motion-heavy activities. The device processes 10 optical channels using an on-device, 15-layer temporally dilated convolutional neural network (~300K parameters) to yield a 1 Hz heart rate output. Crucial to this model's performance was its training on a massive dataset comprising 10,000 hours of data from 962 participants, curated from a broader corpus of controlled and free-living activities. We evaluated the PW2's sensing performance across two independent validation sets: an in-house fitness dataset (229 participants, 250 hours) and an external free-living dataset (27 participants, 1000+ hours). The system achieved 95% Limits of Agreement of -10.34 to 8.66 BPM during exercise and -6.57 to 7.48 BPM during free-living activities, demonstrating substantially tighter error margins than previous Google devices. Finally, we discuss key design lessons, emphasizing that large-scale deep learning was instrumental in fully leveraging multipath PPG hardware over traditional signal processing approaches.
Innovation
Methods, ideas, or system contributions that make the work stand out.
multipath PPG
on-device deep learning
heart rate sensing
temporally dilated CNN
large-scale wearable dataset
🔎 Similar Papers
No similar papers found.
