Abstract
Commercial wearables such as Garmin and Whoop provide users with a daily 'body bat-
tery' score: a numerical indicator of physical and mental readiness. While popular, these
scores are often generated by black-box algorithms, leaving their predictive reliability for
performance or illness detection largely unclear.
The central research question is whether we can build more accurate and transparent
health models than those currently available on the market. To this end, we will also
study innovative Speckle Plethysmography (SPG) sensors. SPG provides biometric sig-
nals such as heart rate, heart rate variability, respiratory rate, and potentially blood
pressure or oxygen saturation, with improved robustness under challenging measurement
conditions.
We will develop statistical and AI-based models that are both predictive and explainable.
These models will be compared to existing commercial scores, and we will investigate the
physiological drivers behind the predictions.
Researcher(s)
Research team(s)
Project type(s)