Method for improving accuracy of pulse rate estimation

1. A method comprising:

receiving, from a sensor, a plurality of samples of a signal, wherein the signal is related to a cardiovascular pulse;

determining a predicted pulse rate based at least on a first set of the plurality of samples;

determining a frequency spectrum of a second set of the plurality of samples, wherein the frequency spectrum comprises a plurality of frequency components, each having a respective magnitude;

determining a frequency of a local maximum of the frequency spectrum;

determining an overall energy in the frequency spectrum;

determining a confidence level of the determined frequency of the local maximum, wherein the determined confidence level is based on at least a magnitude of the local maximum of the frequency spectrum and the determined overall energy in the frequency spectrum; and

determining a pulse rate based on the predicted pulse rate, the determined frequency of the local maximum, and the determined confidence level.

2. The method of claim 1 , wherein determining a pulse rate based on the predicted pulse rate, the determined frequency of the local maximum, and the determined confidence level comprises determining a weighted combination of the predicted pulse rate and the determined frequency based on a weighting factor, wherein the weighting factor is based on the determined confidence level such that a higher determined confidence level causes the pulse rate to be more weighted toward the determined frequency of the local maximum.

3. The method of claim 1 , wherein determining a predicted pulse rate based at least on a first set of the plurality of samples comprises determining a frequency spectrum of the first set of the plurality of samples, wherein the first set of the plurality of samples and the second set of the plurality of samples overlap, and wherein determining a frequency spectrum of the second set of the plurality of samples comprises performing an update on the determined frequency spectrum of the first set of the plurality of samples based on samples of the second set of the plurality of samples that are not part of the first set of the plurality of samples.

4. The method of claim 3 , further comprising:

receiving an artifact signal;

determining a frequency spectrum of the artifact signal; and

filtering the determined frequency spectrum of the second set of the plurality of samples based on the determined frequency spectrum of the artifact signal to generate a plurality of filtered values of the magnitude of the second set of the plurality of samples at the plurality of corresponding different frequencies, wherein determining a frequency of a local maximum of the frequency spectrum comprises determining a frequency of a local maximum of the plurality of filtered values of the magnitude of the second set of the plurality of samples at the plurality of corresponding different frequencies.

5. The method of claim 4 , wherein filtering the determined frequency spectrum of the second set of the plurality of samples based on the determined frequency spectrum of the artifact signal comprises applying a least mean squares filter to a portion of the determined frequency spectrum of the second set of the plurality of samples.

6. The method of claim 1 , wherein the frequency spectrum of the second set of the plurality of samples includes a specified range of frequencies, wherein the specified range of frequencies is between approximately 0.5 Hertz and approximately 3.5 Hertz.

7. The method of claim 1 , wherein determining a frequency of a local maximum of the frequency spectrum comprises:

interpolating the frequency spectrum based on a sinc function; and

determining a frequency of a local maximum of the interpolated frequency spectrum.

8. The method of claim 1 , wherein determining the overall energy in the frequency spectrum comprises determining a sum of the squares of the magnitudes of the frequency components in the frequency spectrum, and wherein determining a confidence level of the determined frequency of the local maximum comprises determining a ratio of the determined sum and the magnitude of a particular frequency component that corresponds to the determined local maximum of the frequency spectrum.

9. The method of claim 1 , further comprising:

receiving an artifact signal; and

filtering the received plurality of samples of the signal based on the artifact signal to generate a plurality of filtered samples, wherein determining a frequency spectrum of a second set of the plurality of samples comprises determining a frequency spectrum of a set of the plurality of filtered samples that correspond to the second set of the plurality of samples.

10. The method of claim 1 wherein receiving a plurality of samples of a signal comprises:

illuminating a portion of subsurface vasculature; and

receiving light emitted from the portion of subsurface vasculature in response to illumination, wherein the plurality of samples of a signal comprise measurements of the received light at a plurality of respective points in time.

11. A system comprising:

a sensor configured to detect a signal, wherein the signal is related to a cardiovascular pulse; and

a controller operably coupled to the sensor, wherein the controller comprises a computing device programmed to perform controller operations comprising:

operating the sensor to obtain a plurality of samples of the signal;

determining a predicted pulse rate based at least on a first set of the plurality of samples;

determining a frequency spectrum of a second set of the plurality of samples, wherein the frequency spectrum comprises a plurality of frequency components, each having a respective magnitude;

determining a frequency of a local maximum of the frequency spectrum;

determining an overall energy in the frequency spectrum;

determining a confidence level of the determined frequency of the local maximum, wherein the determined confidence level is based on at least a magnitude of the local maximum of the frequency spectrum and the determined overall energy in the frequency spectrum; and

determining a pulse rate based on the predicted pulse rate, the determined frequency of the local maximum, and the determined confidence level.

12. The system of claim 11 , wherein the sensor comprises:

a light emitter; and

a light detector, and wherein operating the sensor to obtain a plurality of samples of the signal comprises: (a) operating the light emitter to illuminate a portion of subsurface vasculature and (b) operating the light detector to detect light emitted from the portion of subsurface vasculature in response to illumination at a plurality of points in time, wherein the plurality of samples of the signal comprise measurements of the detected light at respective points in time.

13. The system of claim 11 , further comprising an accelerometer, wherein

the controller operations further comprise:

operating the accelerometer to receive an artifact signal;

determining a frequency spectrum of the artifact signal; and

filtering the determined frequency spectrum of the second set of the plurality of samples based on the determined frequency spectrum of the artifact signal to generate a plurality of filtered values of the magnitude of the second set of the plurality of samples at the plurality of corresponding different frequencies, wherein determining a frequency of a local maximum of the frequency spectrum comprises determining a frequency of a local maximum of the plurality of filtered values of the magnitude of the second set of the plurality of samples at the plurality of corresponding different frequencies.

14. The system of claim 11 , wherein determining a pulse rate based on the predicted pulse rate, the determined frequency of the local maximum, and the determined confidence level comprises determining a weighted combination of the predicted pulse rate and the determined frequency based on a weighting factor, wherein the weighting factor is based on the determined confidence level such that a higher determined confidence level causes the determined pulse rate to be more weighted toward the determined frequency of the local maximum.

15. The system of claim 11 , wherein determining a predicted pulse rate based at least on a first set of the plurality of samples comprises determining a frequency spectrum of the first set of the plurality of samples, wherein the first set of the plurality of samples and the second set of the plurality of samples overlap, and wherein determining a frequency spectrum of the second set of the plurality of samples comprises performing an update on the determined frequency spectrum of the first set of the plurality of samples based on samples of the second set of the plurality of samples that are not part of the first set of the plurality of samples.

16. The system of claim 11 , further comprising an accelerometer, wherein the controller operations further comprise:

operating the accelerometer to receive an artifact signal; and

filtering the received plurality of samples of the signal based on the artifact signal to generate a plurality of filtered samples, wherein determining a frequency spectrum of a second set of the plurality of samples comprises determining a frequency spectrum of a set of the plurality of filtered samples that correspond to the second set of the plurality of samples.

17. The system of claim 11 , wherein the frequency spectrum of the second set of the plurality of samples includes a specified range of frequencies, wherein the specified range of frequencies is between approximately 0.5 Hertz and approximately 3.5 Hertz.

18. The system of claim 11 , wherein determining a frequency of a local maximum of the frequency spectrum comprises:

interpolating the frequency spectrum based on a sinc function; and

determining a frequency of a local maximum of the interpolated frequency spectrum.

19. The system of claim 11 , wherein the system includes a user interface, and wherein the controller operations further comprise:

operating the user interface to provide an indication of the determined pulse rate.

20. The system of claim 11 , wherein the system comprises a body-mountable device.