
Investigators:
Longfei Shangguan, University of Pittsburgh
MassAITC Cohort: Year 4 (Healthy Aging)
Today, smartwatches have emerged as the go-to solution for continuous vital sign monitoring. These devices embed a Photoplethysmography (PPG) sensor which operates by shining a red laser beam into the skin. The laser light reflects back as a function of the blood volume flowing through the veins in the wrist; these reflections are then converted to the user’s heart rate. The smartwatch’s comfortable design allows for day-long usage, facilitating continuous vital sign monitoring for extended durations.
Despite their convenience, existing PPG-based smartwatches are shown to be less accurate in people with dark skin. This discrepancy arises due to the elevated melanin levels in darker skin, causing absorption of the laser light and consequently reducing the signal-to-noise ratio of the measurements.
This research project aims to develop a wearable device for accurate and inclusive user vital sign monitoring based on our recent development of an innovative sensor capable of measuring heart rate and breath by indirectly assessing the density variation of iron within hemoglobin in red blood cells. The measurement of these vital signs from red blood cells requires an external magnetic field to generate a biomagnetic field from the blood vessels. However, the individual-specific factors such as wrist size, fat thickness, and blood vessel dimensions can influence the external magnetic field signal and thus require tuning on a per-user basis. In this project, we will develop a reinforcement learning-based algorithm that can automatically tune the external magnetic field for different users in order to obtain reliable vital sign readings. We will also partner with the Osher Lifelong Learning Institute (OLLI) at the University of Pittsburgh to recruit 50 elder adults for system evaluation.