SYSTEM AND METHOD FOR DETERMINING WHETHER AN ELECTRONIC DEVICE IS LOCATED ON A STATIONARY OR STABLE SURFACE

• US 20200132717A1
• Filed: 10/30/2018
• Published: 04/30/2020
• Est. Priority Date: 10/30/2018
• Status: Active Grant

• Current Assignee: STMicroelectronics SRL (STMicroelectronics NV)
• Original Assignee: STMicroelectronics SRL (STMicroelectronics NV)
• Inventors: Rivolta, Stefano Paolo, Rizzardini, Federico
• Granted Patent: US 11,099,208 B2
• Time in Patent Office: Days
• CPC Class Codes: G01P 15/14 by making use of gyroscopes gyroscopes per se G01C19/00
• CPC Class Codes: G01P 15/18 in two or more dimensions
• CPC Class Codes: G06F 1/3231 Monitoring the presence, absence or movement of users
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

An embodiment system includes: a first motion sensor configured to generate first sensor data indicative of a first type of movement of an electronic device; a first feature detection circuit configured to determine at least one orientation-independent feature based on the first sensor data; and a classifying circuit configured to determine whether or not the electronic device is located on a stationary surface based on the at least one orientation-independent feature.

Claims

1. A system, comprising:

a first motion sensor configured to generate first sensor data indicative of a first type of movement of an electronic device;

a first feature detection circuit configured to determine at least one orientation-independent feature based on the first sensor data; and

a classifying circuit configured to determine whether or not the electronic device is located on a stationary surface based on the at least one orientation-independent feature.

2. The system of claim 1, wherein the first motion sensor comprises an accelerometer of the electronic device.

3. The system of claim 1, wherein the first sensor data comprises a plurality of samples within an acquisition time window, and wherein the first feature detection circuit is configured to:

determine a norm of each of the plurality of samples of the first sensor data within the acquisition time window;

determine, based on the norm of each of the plurality of samples of the first sensor data within the acquisition time window, a mean of the norm within the acquisition time window; and

determine a mean-cross value associated with the acquisition time window, wherein the mean-cross value comprises a number of times the norm of the plurality of samples of the first sensor data within the acquisition time window crosses the mean of the norm within the acquisition time window, and wherein the at least one orientation-independent feature comprises the mean-cross value associated with the acquisition time window.

4. The system of claim 3, wherein the first feature detection circuit is further configured to:

determine a variance of the norm of the plurality of samples of the first sensor data within the acquisition time window, wherein the at least one orientation-independent feature further comprises the variance of the norm of the plurality of samples of the first sensor data within the acquisition time window.

5. The system of claim 3, wherein the acquisition time window is one of a plurality of acquisition time windows that are consecutive and non-overlapping.

6. The system of claim 5, wherein each of the plurality of acquisition time windows has a duration of 1 second.

7. The system of claim 3, wherein the first sensor data comprises 50 samples within the acquisition time window.

8. The system of claim 1, further comprising:

a second motion sensor configured to generate second sensor data indicative of a second type of movement of the electronic device; and

a second feature detection circuit configured to determine at least one further orientation-independent feature based on the second sensor data, wherein the classifying circuit is configured to determine whether the electronic device is located on the stationary surface based on the at least one orientation-independent feature and the at least one further orientation-independent feature.

9. The system of claim 8, wherein the second motion sensor comprises a gyroscope of the electronic device.

10. The system of claim 1, wherein the classifying circuit comprises a machine learning classifier.

11. A method, comprising:

generating, by a first motion sensor of an electronic device, first sensor data over an acquisition time window;

determining, by a first feature detection circuit of the electronic device, at least one first orientation-independent feature for the acquisition time window based on the first sensor data; and

executing, by a classification circuit of the electronic device, a machine learning classification to determine whether or not the electronic device is located on a stationary surface based on the at least one first orientation-independent feature.

12. The method of claim 11, wherein the machine learning classification comprises at least one of a logistic regression classification, a naive Bayes classification, a support vector machine classification, a decision tree classification, a boosted tree classification, a random forest classification, a neural network classification, or a nearest neighbor classification.

13. The method of claim 11, wherein determining, by the first feature detection circuit, the at least one first orientation-independent feature based on the first sensor data comprises:

determining a norm of each sample of the first sensor data within the acquisition time window;

determining a mean of the norm of the first sensor data within the acquisition time window; and

determining a first mean-cross value for the acquisition time window, wherein the first mean-cross value comprises a number of times the norm of the first sensor data within the acquisition time window crosses the mean of the norm of the first sensor data within the acquisition time window, and wherein the at least one first orientation-independent feature comprises the first mean-cross value.

14. The method of claim 13, wherein the first mean-cross value when the electronic device is located on the stationary surface is greater than the first mean-cross value when the electronic device is not located on the stationary surface.

15. The method of claim 11, further comprising:

generating, by a second motion sensor of the electronic device, second sensor data over the acquisition time window; and

determining, by a second feature detection circuit, at least one second orientation-independent feature for the acquisition time window based on the second sensor data, wherein the executing comprises executing the machine learning classification to determine whether or not the electronic device is located on the stationary surface based on the at least one first orientation-independent feature and the at least one second orientation-independent feature.

16. The method of claim 15, wherein determining, by the second feature detection circuit, the at least one second orientation-independent feature based on the second sensor data comprises:

determining a norm of each sample of the second sensor data within the acquisition time window;

determining a mean of the norm of the second sensor data within the acquisition time window; and

determining a second mean-cross value for the acquisition time window, wherein the second mean-cross value comprises a number of times the norm of the second sensor data within the acquisition time window crosses the mean of the norm of the second sensor data within the acquisition time window, and wherein the at least one second orientation-independent feature comprises the second mean-cross value.

17. The method of claim 15, wherein the first motion sensor comprises an accelerometer of the electronic device.

18. The method of claim 15, wherein the second motion sensor comprises a gyroscope of the electronic device.

19. The method of claim 11, wherein executing, by the classification circuit, the machine learning classification comprises assigning a label to the acquisition time window, the label being indicative of whether or not the electronic device is located on the stationary surface.

20. The method of claim 19, further comprising executing, by a meta-classifying circuit, a low pass filtering operation on a plurality of labels output by the classification circuit, the plurality of labels corresponding to consecutive acquisition time windows.

21. An electronic device, comprising:

a detection system, comprising;

an accelerometer configured to generate accelerometer data indicative of a first type of movement of an electronic device;

a first feature detection circuit coupled to an output of the accelerometer and configured to determine at least one orientation-independent feature based on the accelerometer data; and

a classifying circuit configured to determine whether or not the electronic device is located on a stationary surface based on the at least one orientation-independent feature.

22. The electronic device of claim 21, wherein the detection system is implemented as a system-on-chip in the electronic device.

23. The electronic device of claim 21, wherein the detection system is implemented as a microcontroller in the electronic device.

24. The electronic device of claim 21, wherein the detection system is implemented in a micro-electro-mechanical (MEMS) system-in-package in the electronic device.