Time and frequency domain based activity tracking system

Patent metadata

Field	Value
Publication number	US-2016296144-A1
Application number	US-201514681438-A
Country	US
Kind code	A1
Filing date	Apr 8, 2015
Priority date	Apr 29, 2014
Publication date	Oct 13, 2016
Grant date	—

Abstract

A method and apparatus for activity tracking based on time domain and frequency domain processing are disclosed. Embodiments according to the present invention are used to improve the accuracy of activity detection and step counting. The activity tracking starts from sample collection to generate 3-D accelerometer data. By pre-processing, the 3-D accelerometer data is calibrated and filtered. Then, the dominant component is calculated and statistical attributes or features used for activity detection are extracted. The statistical attributes are derived from time domain sensor data, frequency domain transformed data, or both. A classifier is developed using representative training data set. The activity detector determines the current activity status based on the statistical attributes and the classifier. To further refine the activity, post-processing is performed on the activity status.

First claim

1 . A system for activity tracking, the system comprising:
a 3-D accelerometer; a pre-processor module coupled to the 3-D accelerometer; a dominant component computation unit coupled to the pre-processor module; a statistics generator module coupled to the dominant component computation unit and the pre-processor module; and a classifier for detecting an activity based on outputs of the dominant component computation unit and the statistics generator module.

2 . The system of claim 1 , further including a post-processor that is configured to refine an output of the classifier prior to outputting a result of the classifier.

3 . The system of claim 1 , further including a step counter coupled to the classifier and the dominant component computation unit and configured to output number of steps based on an output of the dominant component computation unit to the classifier.

4 . The system of claim 3 , further including a post-processor that is configured to refine an output of the step counter prior to outputting a result of the step counter.

5 . The system of claim 1 , wherein the statistics generator includes a first statistics module and a second statistics module, wherein the first statistics module is configured to generate statistics and frequency domain transformation based on an output of the dominant component computation unit and the second statistics module is configured to generate statistics and frequency domain transformation based on an output of the pre-processor module.

6 . The system of claim 1 , wherein the pre-processor is configured to remove at least one of a direct current (DC) component in a vector sample received from the 3-D accelerometer and a high frequency component from the vector sample.

7 . The system of claim 6 , wherein the dominant component computation unit is configured to determine a dominant component in the vector sample.

8 . The system of claim 7 , wherein the vector sample includes a horizontal component and a vertical component.

9 . The system of claim 8 , wherein the vertical component is proportional to a vector inner product between a mean vector of a plurality of vector samples along three axes and each dynamic vector corresponding to a first difference between each of the plurality of vector samples and the mean vector.

10 . The system of claim 9 , wherein the horizontal component is related to a norm of a second difference between the first difference and the mean vector scaled by the vertical component.

11 . The system of claim 1 , wherein the pre-processor module includes at least one of a level normalization module, a DC notch filter and a low pass filter.

12 . The system of claim 11 , wherein the normalization module is configured to remove a gravity component from the output of the 3-D accelerometer.

13 . A non-transitory computer readable media including programming instruction which when executed by a processor performs an operation, the operation includes:
pre-processing an output of a 3-D accelerometer, wherein the pre-processing includes removing a direct current (DC) component from an output of the 3-D accelerometer; determining a dominant component in the output of the 3-D accelerometer after the pre-processing, wherein the output of the 3-D accelerometer is used to derive a vertical component and a horizontal component, the vertical component is proportional to a vector inner product between a mean vector of a plurality of vector samples along three axes and each dynamic vector corresponding to a first difference between each of the plurality of vector samples and the mean vector, wherein the horizontal component is related to a norm of a second difference between the first difference and the mean vector scaled by the vertical component; and determining an activity type based on statistical data and the dominant component.

14 . The computer readable media of claim 13 , further including refining the activity type based on a step counting, wherein the step counting is calculated based on the dominant component.

15 . The computer readable media of claim 13 , wherein the statistical data includes a first statistical data generated based on the pre-processing and a second statistical data generated based on the determining the dominant component.

16 . The computer readable media of claim 13 , the determining the activity type is further based on cumulative distribution for first values of the statistical data, wherein the cumulative distribution is divided into multiple sections associated with first values of the statistical data and second values of a plurality of statistical features of a particular activity, and wherein each section corresponds to a ratio range of the first values to the second values.

17 . The computer readable media of claim 13 , wherein the multiple sections correspond to five sections.

18 . The computer readable media of claim 17 , wherein the five sections correspond to great than 0 and equal to or less than 5%, great than 5% and equal to or less than 10%, great than 10% and equal to or less than 90%, great than 90% and equal to or less than 95%, and great than 95% and equal to or less than 99%.

19 . A computer readable media including programming instructions which when executed by a processor performs an operation for determining an activity, the operation includes:
receiving raw data from a 3-D accelerometer; filtering the raw data using a low pass filter and a high pass filter and removing a direct current (DC) component from the raw data; calculating a time domain mean value and a time domain energy of the filtered raw data; estimating dominant component in the filtered raw data; and determining the activity based on the estimating and the calculating.

20 . The computer readable media of claim 19 , wherein the operation further includes calculating a mean value prior to the estimating.

Assignees

• Nxp Bv

Inventors

• Gaddam Vasanth
• Zhang Yifeng
• Zhang Jie
• Wu Yuanwei
• Wang Guanqing

Classifications

• G16H50/20

  for computer-aided diagnosis, e.g. based on medical expert systems · CPC title

• H10P50/71

  using masks for conductive or resistive materials · CPC title

• G01C21/20

  Instruments for performing navigational calculations (G01C21/24, G01C21/26 take precedence) · CPC title

• G01P15/00

  Measuring acceleration; Measuring deceleration; Measuring shock, i.e. sudden change of acceleration · CPC title

• A61B5/725

  using specific filters therefor, e.g. Kalman or adaptive filters (specific diagnostics methods using using bioelectric or biomagnetic signals A61B5/316) · CPC title