Context-aware heart rate estimation

• US 9,848,823 B2
• Filed: 07/29/2015
• Issued: 12/26/2017
• Est. Priority Date: 05/29/2014
• Status: Active Grant

• Current Assignee: Apple Inc.
• Original Assignee: Apple Inc.
• Inventors: Raghuram, Karthik Jayaraman, Silva, James M., Kanaris, Alexander, Bhaskaran, Barker N.
• Primary Examiner(s): MANUEL, GEORGE C
• Application Number:
  US14/812,870
• Publication Number:
  US 20160058367A1
• Time in Patent Office: 881 Days
• CPC Class Codes: A61B 5/0205 Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
• CPC Class Codes: A61B 5/02438 with portable devices, e.g. worn by the patient
• CPC Class Codes: A61B 5/1118 Determining activity level
• CPC Class Codes: A61B 5/486 Bio-feedback A61B5/375 takes precedence
• CPC Class Codes: A61B 5/681 Wristwatch-type devices
• CPC Class Codes: A61B 5/721 using a separate sensor to detect motion or using motion information derived from signals other than the physiological signal to be measured
• CPC Class Codes: A61B 5/7246 using correlation, e.g. template matching or determination of similarity
• CPC Class Codes: A61B 5/7278 Artificial waveform generation or derivation, e.g. synthesising signals from measured signals A61B5/327 takes precedence

Abstract

A device can estimate the heart rate of an active user by using a physiological model to refine a “direct” measurement of the user'"'"'s heart rate obtained using a pulse sensor. The physiological model can be based on heart rate response to activity and can be informed by context information, such as the user'"'"'s current activity and/or intensity level as well as user-specific parameters such as age, gender, general fitness level, previous heart rate measurements, etc. The physiological model can be used to predict a heart rate, and the prediction can be used to assess or improve the direct measurement.

Claims

1. A method of determining a heart rate of a user, the method comprising, by a wearable device:

generating, using a pulse sensor of the wearable device, a sequence of heart rate data samples;

determining, using the sequence of heart rate data samples, a direct heart rate estimate;

determining an activity context of the wearable device, the activity context including a physical activity in which the user is engaged and an intensity of the physical activity;

generating, based in part on the activity context and in part on a user-specific parameter, a modeled heart rate estimate, the modeled heart rate estimate being based on a physiological model of heart rate response to the activity context and generated independently of the sequence of heart rate data samples; and

determining a final heart rate estimate based on the direct heart rate estimate and the modeled heart rate estimate.

2. The method of claim 1 wherein determining the direct heart rate estimate includes:

determining, using a motion sensor of the wearable device, a motion-related noise component; and

subtracting the motion-related noise component from the heart rate data samples.

3. The method of claim 1 wherein the pulse sensor includes a photoplethysmographic sensor and wherein determining the direct heart rate estimate includes:

measuring a dark channel signal for the photoplethysmographic sensor; and

subtracting the dark channel signal from the heart rate data samples.

4. The method of claim 1 wherein determining the direct heart rate estimate includes:

generating a frequency spectrum from the heart rate data samples;

comparing the generated frequency spectrum to each of a plurality of template spectra, each template spectrum corresponding to a different candidate heart rate; and

identifying a best matching template spectrum based on the comparing,

wherein the direct heart rate estimate is based on the best-matching template spectrum.

5. The method of claim 4 further comprising:

generating a confidence score for the direct heart rate estimate, the confidence score based at least in part on how well the generated frequency spectrum matches the best-matching template spectrum,

wherein determining the final heart rate estimate is based in part on the confidence score.

6. The method of claim 1 wherein the activity context includes an activity identifier and wherein generating the modeled heart rate estimate includes:

determining a starting time based on a most recent change in the activity identifier;

establishing a starting heart rate based on a heart rate determined prior to the starting time;

determining an equilibrium heart rate based at least in part on a current activity indicated by the activity identifier; and

applying a filter function to the starting heart rate and the equilibrium heart rate based on a current time and the starting time.

7. The method of claim 6 wherein the equilibrium heart rate is determined based in part on the current activity and in part on a user-specific parameter.

8. The method of claim 7 wherein the user-specific parameter includes a parameter indicating a demographic characteristic of the user.

9. The method of claim 7 wherein the user-specific parameter includes a parameter indicating a fitness level of the user.

10. The method of claim 7 wherein the user-specific parameter includes a previously determined user-specific equilibrium heart rate associated with the current activity.

11. The method of claim 1 wherein generating the modeled heart rate estimate includes generating a probability mass function indicating a probability of each of a plurality of possible heart rates,

wherein determining the final heart rate estimate is based in part on the probability mass function.

12. A device comprising:

a pulse sensor;

a motion sensor; and

a processor coupled to the pulse sensor and the motion sensor, the processor configured to;

obtain a sequence of heart rate data samples from the pulse sensor;

determine, based on the sequence of heart rate data samples, a direct heart rate estimate;

determine, based at least in part on data from the motion sensor, an activity context of the device, the activity context including a physical activity in which a user is engaged and an intensity of the physical activity;

generate, based in part on the activity context of the device and in part on a user-specific parameter, a modeled heart rate estimate, the modeled heart rate estimate being based on a physiological model of heart rate response to the activity context and generated independently of the sequence of heart rate data samples; and

determine a final heart rate estimate based on the direct heart rate estimate and the modeled heart rate estimate.

13. The device of claim 12 wherein the pulse sensor includes a photoplethysmographic (PPG) sensor.

14. The device of claim 12 wherein the motion sensor includes one or more of an accelerometer or a gyroscope.

15. The device of claim 12 further comprising a user interface coupled to the processor and operable to present the final heart rate estimate to the user.

16. The device of claim 12 wherein the physiological model includes a filter function to model a heart rate transition from a staring heart rate to an equilibrium heart rate associated with a current activity context of the device.

17. A computer-readable storage medium having stored thereon program instructions that, when executed by a processor, cause the processor to perform a method comprising:

obtaining a sequence of heart rate data samples for a user;

determining, using the sequence of heart rate data samples, a direct heart rate estimate;

determining an activity context for the user, the activity context including a physical activity in which the user is engaged and an intensity of the physical activity;

generating, based in part on the activity context and in part on a user-specific parameter, a modeled heart rate estimate, the modeled heart rate estimate being based on a physiological model of heart rate response to the activity context and generated independently of the sequence of heart rate data samples; and

determining a final heart rate estimate based at least in part on the direct heart rate estimate and the modeled heart rate estimate.

18. The computer-readable storage medium of claim 17 wherein the method further comprises:

updating the user-specific parameter based on the final heart rate estimate.

19. The computer-readable storage medium of claim 17 wherein the method further comprises:

determining a confidence score for the direct heart rate estimate; and

determining a probability mass function for the modeled heart rate estimate,

wherein the final heart rate estimate is based in part on the confidence score and the probability mass function.