Detection of biking, walking, and running

1. A method comprising:

receiving motion data collected by a motion sensor worn by a user during an activity;

determining a magnitude of the motion data by taking the magnitude of an acceleration signal, applying a high-pass filter to the acceleration signal, and integrating the standard deviation of the high-passed acceleration signal over a period of time;

determining a frequency of the motion data; and

identifying the activity based on the magnitude and the frequency of the motion data.

2. The method of claim 1 further comprising:

determining the frequency of the motion data by determining a cadence of steps taken by the user;

responsive to the cadence exceeding a cadence threshold and the magnitude of the motion data exceeding a magnitude threshold, identifying the activity as running; and

responsive to at least one of the cadence and magnitude being less than a corresponding threshold, identifying the activity as walking.

3. The method of claim 2 further comprising:

after identifying the activity as running, detecting at least one of the cadence falling below the cadence threshold and the magnitude of the motion data falling below the magnitude threshold; and

responsive to the detection, identifying the activity as walking.

4. The method of claim 2 , wherein the cadence is determined using a period of the magnitude of the motion data.

5. The method of claim 1 further comprising:

determining an orientation of the wearable device; and

identifying the activity as bicycling based on the magnitude and the frequency of the motion data and the orientation of the wearable device.

6. The method of claim 5 further comprising:

determining a dominant frequency of a first frequency range of the motion data and a dominant frequency of a second frequency range of the motion data.

7. The method of claim 6 further comprising:

applying a trained classifier to the magnitude of the motion data, the orientation of the wearable device, the dominant frequency of the first frequency range, and the dominant frequency of the second frequency range, the classifier generating a score indicating a likelihood that the user is bicycling; and

responsive to the score exceeding a threshold, identifying the activity as bicycling.

8. The method of claim 1 further comprising:

automatically starting a timer to time the activity responsive to identifying the activity.

9. The method of claim 1 further comprising:

determining a caloric expenditure of the user during the activity based at least in part on the identified activity.

10. A computer program product comprising a tangible non-transitory computer readable storage medium containing computer program instructions for identifying an activity in which a user is engaged, the computer program instructions executable by a processor to:

receive motion data of the user collected by a wearable device worn by the user during the activity;

determine a magnitude of the motion data by taking the magnitude of an acceleration signal, applying a high-pass filter to the acceleration signal, and integrating the standard deviation of the high-passed acceleration signal over a period of time;

determine a frequency of the motion data; and

identify the activity based on the magnitude and the frequency of the motion data.

11. The computer program product of claim 10 , wherein the computer program instructions executable by the processor to determine the frequency of the motion data comprise computer program instructions executable by the processor to determine a cadence of steps taken by the user, and wherein the computer program instructions executable by the processor to identify the activity comprise instructions executable by the processor to:

responsive to the cadence exceeding a cadence threshold and the magnitude of the motion data exceeding a magnitude threshold, identify the activity as running; and

responsive to at least one of the cadence and magnitude being less than a corresponding threshold, identify the activity as walking;

after identifying the activity as running, detect at least one of the cadence falling below the cadence threshold and the magnitude of the motion data falling below the magnitude threshold; and

responsive to the detection, identify the activity as walking.

12. The computer program product of claim 10 , wherein the computer program instructions executable by the processor to identify the activity comprise computer program instructions executable by the processor to:

determine an orientation of the wearable device; and

identify the activity as bicycling based on the magnitude and the frequency of the motion data and the orientation of the wearable device.

13. The computer program product of claim 12 , wherein the computer program instructions executable by the processor to determine the frequency of the motion data comprise computer program instructions executable by the processor to:

determine a dominant frequency of a first frequency range of the motion data and a dominant frequency of a second frequency range of the motion data.

14. The computer program product of claim 13 , wherein the computer program instructions executable by the processor to identify the activity as bicycling comprise computer program instructions executable by the processor to:

apply a trained classifier to the magnitude of the motion data, the orientation of the wearable device, the dominant frequency of the first frequency range, and the dominant frequency of the second frequency range, the classifier generating a score indicating a likelihood that the user is bicycling; and

responsive to the score exceeding a threshold, identify the activity as bicycling.