Method and apparatus for sensor based pedestrian motion detection in hand-held devices

• US 8,849,605 B2
• Filed: 08/23/2011
• Issued: 09/30/2014
• Est. Priority Date: 08/23/2011
• Status: Expired due to Fees

• Current Assignee: Qualcomm, Inc.
• Original Assignee: Qualcomm, Inc.
• Inventors: Lakhzouri, Abdelmonaem, Viskari, Joonas V. T.
• Primary Examiner(s): Breene, John
• Assistant Examiner(s): Zhang, Ruihua
• Application Number:
  US13/215,580
• Publication Number:
  US 20130054181A1
• Time in Patent Office: 1,134 Days
• Field of Search: 702/85, 702/141, 73/13.7
• US Class Current: 702/141
• CPC Class Codes: G01C 21/20 Instruments for performing navigational calculations G01C21/24, G01C21/26 take precedence
• CPC Class Codes: G01C 22/006 Pedometers

Abstract

A method and apparatus are described for sensor based detection of pedestrian motion. Based on a 3-axis accelerometer, the apparatus may differentiate between walking, running, standing still, or any random movement that the user may perform. The method may comprise the steps of performing a time domain analysis and a frequency domain analysis. The time domain analysis may be based on a Teager-Kaiser Energy Operator. The frequency domain analysis may be based on a fast Fourier transform. A apparatus for detecting pedestrian motion may comprises an accelerometer, an operator, a Teager-Kaiser Energy Operator, a first peak detection, a second peak detection, a buffer, a fast Fourier transform, a memory and a look-up table. The apparatus may be a hand held device.

Claims

1. A computer-implemented method of detecting motion of a pedestrian, the method performed by one or more processors and comprising the steps of:

determining a magnitude of acceleration;

performing a time domain analysis on the acceleration magnitude to generate a time-based representation of the acceleration;

performing a frequency domain analysis on the acceleration magnitude to generate a frequency-based representation of the acceleration; and

determining whether the pedestrian is in motion in response to the time-based representation and the frequency-based representation, wherein determining whether the pedestrian is in motion comprises;

determining whether a number of first peaks of the time-based representation exceed a first threshold;

determining whether a number of second peaks of the frequency-based representation exceed a second threshold; and

comparing the first peaks and one of the second peaks with predetermined values.

2. The method of claim 1, wherein the time domain analysis is performed using a Teager-Kaiser Energy Operator.

3. The method of claim 1, wherein the frequency domain analysis is performed using a fast Fourier transform.

4. The method of claim 3, wherein the acceleration magnitude is a rectangular shape with a periodicity.

6. The method of claim 1, wherein the pedestrian motion comprises a hand motion and a step motion.

7. The method of claim 6, wherein the step motion comprises running and walking.

8. The method of claim 6, wherein the hand motion comprises shaking, waving, and gesturing.

5. A computer-implemented method of detecting motion of a pedestrian, the method performed by one or more processors and comprising the steps of:

determining a magnitude of acceleration;

performing a time domain analysis on the acceleration magnitude to generate a time-based representation of the acceleration;

performing a frequency domain analysis on the acceleration magnitude to generate a frequency-based representation of the acceleration; and

determining whether the pedestrian is in motion in response to the time-based representation and the frequency-based representation, wherein determining whether the pedestrian is in motion further comprises the steps of;

detecting a number of first peaks in the time-based representation;

detecting a number of second peaks in the frequency-based representation; and

determining whether the number of first peaks or second peaks indicates motion of the pedestrian, wherein determining whether the number of first peaks or second peaks indicates motion of the pedestrian comprises;

determining a first time value associated with one of the first peaks;

computing a first frequency value from the first time value;

determining a second frequency value associated with one of the second peaks; and

comparing the first and second frequency values with a predetermined frequency value stored in a look-up table.

9. A computer-implemented method of detecting motion of a pedestrian, the method performed by one or more processors and comprising the steps of:

determining a magnitude of acceleration;

performing a time domain analysis on the acceleration magnitude to generate a time-based representation of the acceleration;

performing a frequency domain analysis on the acceleration magnitude to generate a frequency-based representation of the acceleration; and

determining whether the pedestrian is in motion in response to the time-based representation and the frequency-based representation, wherein determining whether the pedestrian is in motion further comprises the steps of;

determining whether a number of first peaks of the time-based representation exceed a first threshold;

if the first peaks do not exceed the first threshold, indicating no pedestrian motion;

if the first peaks exceed the first threshold, computing a first frequency value from a pair of first peak detection values;

determining whether a number of second peaks of the frequency-based representation exceed a second threshold;

if the second peaks do not exceed the second threshold, indicating no pedestrian motion;

if the second peaks exceed the second threshold, generating a second peak detection value;

comparing the first frequency value and the second peak detection value with predetermined frequency values stored in a look-up table; and

selectively indicating pedestrian motion in response to the comparing.

10. An apparatus for detecting motion of a pedestrian, the pedestrian motion including hand motion and step motion, the apparatus comprising:

an accelerometer having an output to generate a magnitude of acceleration;

an energy operator, coupled to the accelerometer, to generate a time-based representation of the acceleration;

an estimator, coupled to the accelerometer, to generate a frequency-based representation of the acceleration;

a first peak detection circuit, coupled to the energy operator, to determine whether a number of first peaks of the time-based representation exceed a first threshold;

a second peak detection circuit, coupled to the estimator, to determine whether a number of second peaks of the frequency-based representation exceed a second threshold; and

a look-up table, coupled to the first and second peak detection circuits, to compare the first peaks and one of the second peaks with predetermined values stored in the look-up table.

11. The apparatus of claim 10, wherein the look-up table is to generate an output signal indicating the pedestrian motion.

12. The apparatus of claim 10, wherein the energy operator comprises a Teager-Kaiser Energy Operator.

13. The apparatus of claim 10, wherein the estimator comprises a buffer and fast Fourier transform.

14. The apparatus of claim 10, wherein the first peak detection circuit is to indicate no pedestrian motion if the first peaks do not exceed the first threshold.

15. The apparatus of claim 10, wherein the second peak detection circuit is to indicate no pedestrian motion if the second peaks do not exceed the second threshold.

16. The apparatus of claim 10, wherein the magnitude of acceleration comprises a square root of the sum of a_x²₊a_y²+a_z², wherein a_x, a_y, and a_z comprise x, y, and z components of the accelerometer.

17. The apparatus of claim 10, wherein the step motion comprises running and walking.

18. The apparatus of claim 10, wherein the hand motion comprises shaking, waving, and gesturing.

19. The apparatus of claim 10, wherein the apparatus is a hand-held device.