Systems and methods for gating a step count

1. A system comprising:

one or more processors; and

memory coupled to the one or more processors, the memory encoded with a set of instructions configured to perform a process comprising:

obtaining acceleration data for an observation window of an accelerometer;

determining whether to perform step counting by at least:

inputting two or more characteristics of the acceleration data into a decision tree to determine activity occurring within the observation window, wherein the two or more characteristics include an L1 Difference Norm, an L2 Difference Norm, a Fast Fourier Transform (FFT) Peak Frequency, an FFT Spectral Entropy, and FFT Total Energy;

determining, by a decision tree, a probability for each of one or more activities determined to be occurring with the observation window based at least on the two or more characteristics;

determining, by the decision tree, an activity with the greatest probability of occurring within the observation window based on the determined probabilities for the one or more activities;

assigning a first class to the observation window when the determined activity with the greatest probability is associated with consecutive steps; and

assigning a second class to the observation window when the determined activity with the greatest probability is not associated with consecutive steps,

determining a step count for the observation window utilizing frequency analysis when the first class is assigned to the observation window; and

stopping processing of the acceleration data for the observation window without determining the step count for the observation window when the second class is assigned to the observation window.

2. The system of claim 1 , wherein the first class corresponds to low activity or high activity.

3. The system of claim 1 , wherein the frequency analysis includes integrating a dominant frequency in the observation window over a width of the observation window.

4. The system of claim 3 , wherein the width of the observation window is 5.12 seconds.

5. The system of claim 1 , wherein the process further comprises:

obtaining additional acceleration data for another observation window of the accelerometer;

inputting two or more characteristics of the additional acceleration data into the decision tree to determine whether to perform the step counting;

determining, by the decision tree, that additional activity within the another observation window is not associated with consecutive steps;

assigning the second class to the another observation window based on the determination that the additional activity is not associated with the consecutive steps; and

incrementing a count of the second class in a database and stopping processing of the additional acceleration data for the another observation window without determining a step count for the another observation window.

6. The system of claim 1 , wherein the process further comprises incrementing a count of the first class in a database and adding the determined step count for the observation window to a total step count stored in the database.

7. The system of claim 6 , further comprising a continuous glucose monitoring device and the accelerometer.

8. The system of claim 7 , wherein the accelerometer includes a 12.5 Hz sampling rate, the observation window has a width of 5.12 seconds, and there are 64 samples within the observation window.

9. The system of claim 8 , wherein the process further comprises normalizing magnitudes of the 64 samples to be zero mean, and calculating a vector X, wherein the vector X is used to calculate the two or more characteristics of the acceleration data.

10. The system of claim 7 , wherein the process further comprises reporting a glucose reading with a contextual label based on the count of the first class and the total step count in the database.

11. The system of claim 7 , wherein the process further comprises correcting a glucose reading based on the count of the first class and the total step count in the database.

12. A non-transitory computer readable storage medium storing instructions that, when executed by one or more processors of a computing system, cause the computing system to perform operations comprising:

obtaining acceleration data for an observation window of an accelerometer;

determining whether to perform step counting by at least:

inputting two or more characteristics of the acceleration data into a decision tree to determine activity occurring within the observation window, wherein the two or more characteristics include an L1 Difference Norm, an L2 Difference Norm, a Fast Fourier Transform (FFT) Peak Frequency, an FFT Spectral Entropy, and FFT Total Energy;

determining, by a decision tree, a probability for each of one or more activities determined to be occurring with the observation window based at least on the two or more characteristics;

determining, by the decision tree, an activity with the greatest probability of occurring within the observation window based on the determined probabilities for the one or more activities;

assigning a first class to the observation window when the determined activity with the greatest probability is associated with consecutive steps; and

assigning a second class to the observation window when the determined activity with the greatest probability is not associated with consecutive steps,

determining a step count for the observation window utilizing frequency analysis when the first class is assigned to the observation window; and

stopping processing of the acceleration data for the observation window without determining the step count for the observation window when the second class is assigned to the observation window.

13. The non-transitory computer readable storage medium of claim 12 , wherein the operations further comprise:

obtaining additional acceleration data for another observation window of the accelerometer;

inputting two or more characteristics of the additional acceleration data into the decision tree to determine whether to perform the step counting;

determining, by the decision tree, that additional activity within the another observation window is not associated with the consecutive steps;

assigning the second class to the another observation window based on the determination that the activity is associated with the consecutive steps; and

incrementing a count of the second class in a database and stopping processing of the additional acceleration data for the another observation window without determining a step count for the another observation window.

14. The non-transitory computer readable storage medium of claim 13 , wherein the first class corresponds to low activity or high activity, and the second class corresponds to no activity.

15. The non-transitory computer readable storage medium of claim 14 , wherein the operations further comprise:

normalizing magnitudes of the acceleration data to be zero mean, and calculating a vector X for the acceleration data, wherein the vector X is used to calculate the two or more characteristics of the acceleration data; and

calculating a Fast Fourier Transform (FFT) for the vector X of the acceleration data and the additional acceleration data, respectively.

16. The non-transitory computer readable storage medium of claim 15 , wherein:

the performing the step count comprises detecting and adjusting for harmonics by scaling a magnitude of the FFT for the acceleration data; and

the detecting and adjusting for the harmonics comprises selecting a binary spectrum weight function for the FFT of the acceleration data.

17. A method for monitoring activity comprising:

obtaining, at a computer system, acceleration data for an observation window of an accelerometer;

determining whether to perform step counting by at least:

inputting two or more characteristics of the acceleration data into a decision tree to determine activity occurring within the observation window, wherein the two or more characteristics include an L1 Difference Norm, an L2 Difference Norm, a Fast Fourier Transform (FFT) Peak Frequency, an FFT Spectral Entropy, and FFT Total Energy;

determining, by a decision tree, a probability for each of one or more activities determined to be occurring with the observation window based at least on the two or more characteristics;

determining, by the decision tree, an activity with the greatest probability of occurring within the observation window based on the determined probabilities for the one or more activities;

assigning a first class to the observation window when the determined activity with the greatest probability is associated with consecutive steps; and

assigning a second class to the observation window when the determined activity with the greatest probability is not associated with consecutive steps,

determining, at the computer system, a step count for the observation window utilizing frequency analysis when the first class is assigned to the observation window; and

stopping, at the computer system, processing of the acceleration data for the observation window without determining the step count for the observation window when the second class is assigned to the observation window.

18. The method of claim 17 , further comprising:

obtaining additional acceleration data for another observation window of the accelerometer;

inputting two or more characteristics of the additional acceleration data into the decision tree to determine whether to perform the step counting;

determining, by the decision tree, that additional activity within the another observation window is not associated with the consecutive steps;

assigning the second class to the another observation window based on the determination that the activity is associated with the consecutive steps; and

incrementing a count of the second class in a database and stopping processing of the additional acceleration data for the another observation window without determining a step count for the another observation window.

19. The method of claim 18 , wherein the first class corresponds to low activity or high activity, and the second class corresponds to no activity.

20. The system of claim 1 , wherein the system is configured to be worn by a user.