Wrist-detection algorithm
Algorithms for detecting whether a device is properly secured to a user's skin are described. The operation of a device, such as a wearable device, can be adjusted based on whether the device is properly secured to a user's skin (e.g., on-wrist) or not properly secured to the user's skin (e.g., off-wrist). For example, certain functions can be disabled for power-saving, security or other purposes if the device is off-wrist. In order to avoid falsely identifying the device as off-wrist or on-wrist, algorithms for detecting whether the device is on-wrist or off-wrist can calculate one or more variances based on signals measured by a light sensor and compare the one or more variances with one or more thresholds. Comparing the one or more variances to the one or more threshold can improve the accuracy of wrist-detection algorithms.
1. A device comprising:
a sensor comprising a light emitter and a light detector configured to generate signal samples, wherein each signal sample includes a first signal generated when the light emitter is on and a second signal generated when the light emitter is off; and
processing circuitry capable of:
performing one or more variance calculations for each signal sample;
predicting a condition of the device based on the one or more variance calculations for each signal sample; and
determining that the device is not secured to an object when the predictions of the condition of the device for the signal samples meet a threshold.
2. The device of claim 1 , wherein the one or more variance calculations comprises a first variance calculation of a variance of a difference between the first signal and the second signal of the signal sample.
3. The device of claim 2 , wherein predicting the condition of the device based on the one or more variance calculations for each signal sample comprises, predicting a first condition of the device when the first variance calculation is below a first variance threshold and predicting a second condition of the device when the first variance calculation meets or exceeds the first variance threshold.
4. The device of claim 2 , wherein the one or more variance calculations comprises a second variance calculation of a difference between a variance of the first signal of the signal sample and a variance of the second signal of the signal sample.
5. The device of claim 4 , wherein predicting the condition of the device based on the one or more variance calculations for each signal sample comprises:
predicting a first condition of the device when the first variance calculation is below a first variance threshold;
predicting the first condition of the device when the first variance calculation meets or exceeds the first variance threshold, the first variance calculation is below a second variance threshold, and the second variance calculation is below a third variance threshold;
predicting a second condition of the device when the first variance calculation meets or exceeds the first variance threshold and the first variance calculation meets or exceeds the second variance threshold; and
predicting the second condition of the device when the first variance calculation meets or exceeds the first variance threshold, the first variance calculation is below the second variance threshold, and the second variance calculation meets or exceeds the third variance threshold.
6. The device of claim 4 , wherein predicting the condition of the device based on the one or more variance calculations for each signal sample comprises:
predicting a first condition of the device when the first variance calculation is below a first variance threshold and the first variance calculation is below a third variance threshold;
predicting the first condition of the device when the first variance calculation is below the first variance threshold, the first variance calculation meets or exceeds the third variance threshold, and the second variance calculation is below a fourth variance threshold;
predicting a second condition of the device when the first variance calculation meets or exceeds the first variance threshold and the first variance calculation meets or exceeds the second variance threshold; and
predicting the second condition of the device when the first variance calculation is below the first variance threshold, the first variance calculation meets or exceeds the third variance threshold, and the second variance calculation meets or exceeds the fourth variance threshold.
7. The device of claim 1 , wherein predicting the condition of the device based on the one or more variance calculations for each signal sample comprises predicting a first condition of the device or predicting a second condition of the device, wherein the first condition corresponds to a prediction that the device is not secured to the object and wherein the second condition corresponds to a prediction that the device is secured to the object.
8. The device of claim 7 , wherein determining that the device is not secured to the object when the predictions of the condition of the device for the signal samples meet the threshold comprises predicting the first condition for a threshold number of consecutive predictions.
9. A method for determining that a device is not secured to an object, the method comprising:
generating a plurality of signal samples, wherein each signal sample includes a first signal generated when a light emitter is on and a second signal generated when the light emitter is off;
performing one or more variance calculations for each signal sample;
predicting a condition of the device based on the one or more variance calculations for each signal sample; and
determining that the device is not secured to the object when the predictions of the condition of the device for the signal samples meet a threshold.
10. The method of claim 9 , wherein the one or more variance calculations comprises a first variance calculation of a variance of a difference between the first signal and the second signal of the signal sample.
11. The method of claim 10 , wherein predicting the condition of the device based on the one or more variance calculations for each signal sample comprises, predicting a first condition of the device when the first variance calculation is below a first variance threshold and predicting a second condition of the device when the first variance calculation meets or exceeds the first variance threshold.
12. The method of claim 10 , wherein the one or more variance calculations comprises a second variance calculation of a difference between a variance of the first signal of the signal sample and a variance of the second signal of the signal sample.
13. The method of claim 12 , wherein predicting the condition of the device based on the one or more variance calculations for each signal sample comprises:
predicting a first condition of the device when the first variance calculation is below a first variance threshold;
predicting the first condition of the device when the first variance calculation meets or exceeds the first variance threshold, the first variance calculation is below a second variance threshold, and the second variance calculation is below a third variance threshold;
predicting a second condition of the device when the first variance calculation meets or exceeds the first variance threshold and the first variance calculation meets or exceeds the second variance threshold; and
predicting the second condition of the device when the first variance calculation meets or exceeds the first variance threshold, the first variance calculation is below the second variance threshold, and the second variance calculation meets or exceeds the third variance threshold.
14. The method of claim 12 , wherein predicting the condition of the device based on the one or more variance calculations for each signal sample comprises:
predicting a first condition of the device when the first variance calculation is below a first variance threshold and the first variance calculation is below a third variance threshold;
predicting the first condition of the device when the first variance calculation is below the first variance threshold, the first variance calculation meets or exceeds the third variance threshold, and the second variance calculation is below a fourth variance threshold;
predicting a second condition of the device when the first variance calculation meets or exceeds the first variance threshold and the first variance calculation meets or exceeds the second variance threshold; and
predicting the second condition of the device when the first variance calculation is below the first variance threshold, the first variance calculation meets or exceeds the third variance threshold, and the second variance calculation meets or exceeds the fourth variance threshold.
15. The method of claim 9 , wherein predicting the condition of the device based on the one or more variance calculations for each signal sample comprises predicting a first condition of the device or predicting a second condition of the device, wherein the first condition corresponds to a prediction that the device is not secured to the object and wherein the second condition corresponds to a prediction that the device is secured to the object.
16. The method of claim 15 , wherein determining that the device is not secured to the object when the predictions of the condition of the device for the signal samples meet the threshold comprises predicting the first condition for a threshold number of consecutive predictions.
17. A non-transitory computer readable storage medium, the computer readable medium containing instructions that, when executed, perform a method for operating an electronic device, the electronic device including a processor and a sensor comprising a light emitter and a light detector, the method comprising:
generating a plurality of signal samples, wherein each signal sample includes a first signal generated when the light emitter is on and a second signal generated when the light emitter is off;
performing one or more variance calculations for each signal sample;
predicting a condition of the device based on the one or more variance calculations for each signal sample; and
determining that the device is not secured to an object when the predictions of the condition of the device for the signal samples meet a threshold.
18. The non-transitory computer readable storage medium of claim 17 , wherein the one or more variance calculations comprises a first variance calculation of a variance of a difference between the first signal and the second signal of the signal sample.
19. The non-transitory computer readable storage medium of claim 18 , wherein predicting the condition of the device based on the one or more variance calculations for each signal sample comprises, predicting a first condition of the device when the first variance calculation is below a first variance threshold and predicting a second condition of the device when the first variance calculation meets or exceeds the first variance threshold.
20. The non-transitory computer readable storage medium of claim 18 , wherein the one or more variance calculations comprises a second variance calculation of a difference between a variance of the first signal of the signal sample and a variance of the second signal of the signal sample.
21. The non-transitory computer readable storage medium of claim 20 , wherein predicting the condition of the device based on the one or more variance calculations for each signal sample comprises:
predicting a first condition of the device when the first variance calculation is below a first variance threshold;
predicting the first condition of the device when the first variance calculation meets or exceeds the first variance threshold, the first variance calculation is below a second variance threshold, and the second variance calculation is below a third variance threshold;
predicting a second condition of the device when the first variance calculation meets or exceeds the first variance threshold and the first variance calculation meets or exceeds the second variance threshold; and
predicting the second condition of the device when the first variance calculation meets or exceeds the first variance threshold, the first variance calculation is below the second variance threshold, and the second variance calculation meets or exceeds the third variance threshold.
22. The non-transitory computer readable storage medium of claim 20 , wherein predicting the condition of the device based on the one or more variance calculations for each signal sample comprises:
predicting a first condition of the device when the first variance calculation is below a first variance threshold and the first variance calculation is below a third variance threshold;
predicting the first condition of the device when the first variance calculation is below the first variance threshold, the first variance calculation meets or exceeds the third variance threshold, and the second variance calculation is below a fourth variance threshold;
predicting a second condition of the device when the first variance calculation meets or exceeds the first variance threshold and the first variance calculation meets or exceeds the second variance threshold; and
predicting the second condition of the device when the first variance calculation is below the first variance threshold, the first variance calculation meets or exceeds the third variance threshold, and the second variance calculation meets or exceeds the fourth variance threshold.
23. The non-transitory computer readable storage medium of claim 17 , wherein predicting the condition of the device based on the one or more variance calculations for each signal sample comprises predicting a first condition of the device, wherein the first condition corresponds to a prediction that the device is not secured to the object and wherein the second condition corresponds to a prediction that the device is secured to the object.
24. The non-transitory computer readable storage medium of claim 23 , wherein determining that the device is not secured to the object when the predictions of the condition of the device for the signal samples meet the threshold comprises predicting the first condition for a threshold number of consecutive predictions.