Wearable health monitors and methods of monitoring health
Wearable technologies, such as wearable health monitors, and methods of use are provided. In some embodiments, the wearable technology can be worn at the wrist of an individual and can use an accelerometer, pulse oximeter, and electrocardiogram to measure heart rate, oxygen saturation, blood pressure, pulse wave velocity, and activity. This information can then be provided to the individual. The individual can alter their behaviors and relationships with their own health by using features such as notifications and auto-tagging to better understand their own stress, diet, sleep, and exercise levels over various time periods and subsequently make appropriate behavioral changes.
1 . An apparatus, comprising:
a strap configured to be worn on a first limb of a user;
a pulse oximeter disposed on at least one printed circuit board (PCB) which is coupled to the strap, the pulse oximeter comprising:
at least one light emitting diode (LED) configured to be oriented towards the first limb of the user when the strap is worn by the user and to emit, through openings of the strap and into the first limb of the user, a first type of light having a first peak wavelength and a second type of light having a second peak wavelength; and
a light sensor configured to be oriented towards the first limb of the user when the strap is worn by the user and comprising respective components configured to sense, from the first limb of the user, the first type of light and the second type of light;
an electrocardiogram sensor coupled to the strap, the electrocardiogram sensor comprising:
a first electrode configured to make electrical contact with the first limb of the user when the strap is worn by the user; and
a second electrode configured to be oriented such that the user may make electrical contact with the second electrode using a second limb of the user when the strap is worn by the user; and
a computing system, configured to:
receive data from the pulse oximeter and the electrocardiogram sensor;
determine, using data from the pulse oximeter, an amount of oxygen in blood of the user based on a comparison of:
a first reading sensed by the light sensor corresponding to an amount of the first type of light absorbed by the blood; and
a second reading sensed by the light sensor corresponding to an amount of the second type of light absorbed by the blood;
determine, using data from the pulse oximeter, a pulse onset timing of the user;
determine, using data from the electrocardiogram sensor, an R-wave timing of an electrocardiogram signal of the user;
determine, using the pulse onset timing and the R-wave timing, a pulse wave velocity of the user;
determine, using the pulse wave velocity, a pulse transit time of the user;
input the pulse transit time and the amount of oxygen in the blood into a machine learning model trained, using supervised learning based on a dataset including pulse transit time data and blood oxygenation data, to produce an output indicative of blood pressure;
receive, from the machine learning model, estimates of:
systolic blood pressure of the user; and
diastolic blood pressure of the user; and
output the estimates to the user.
2 . The apparatus of claim 1 , wherein:
the first peak wavelength is about 660 nm; and
the second peak wavelength is about 940 nm.
3 . The apparatus of claim 1 , wherein the machine learning model comprises an artificial neural network (ANN).
4 . The apparatus of claim 3 , wherein a final layer of the ANN comprises a collection of neurons with an output representing a range for at least one of the estimates of systolic blood pressure and diastolic blood pressure.
5 . The apparatus of claim 3 , wherein a final layer of the ANN comprises a single neuron with an output that maps to at least one of the estimates of systolic blood pressure and diastolic blood pressure.
6 . The apparatus of claim 1 , wherein the machine learning model comprises a polynomial regression analysis, a support vector machine, or a combination thereof.