Method for step detection and gait direction estimation

Patent metadata

Field	Value
Publication number	US-10215587-B2
Application number	US-201313791443-A
Country	US
Kind code	B2
Filing date	Mar 8, 2013
Priority date	May 18, 2012
Publication date	Feb 26, 2019
Grant date	Feb 26, 2019

Abstract

A method for detecting a human's steps and estimating the horizontal translation direction and scaling of the resulting motion relative to an inertial sensor is described. When a pedestrian takes a sequence of steps the displacement can be decomposed into a sequence of rotations and translations over each step. A translation is the change in the location of pedestrian's center of mass and a rotation is the change along z-axis of the pedestrian's orientation. A translation can be described by a vector and a rotation by an angle.

First claim

What is claimed:

1. A computer-implemented method for detecting the steps of a person and estimating the person's three-dimensional (3D) movement to track a location of the person, comprising:
collecting accelerometer data from a worn or person carried device that includes an accelerometer in an unknown tilted orientation relative to a ground frame as the person moves around a physical location, the accelerometer data being indicative of the person's 3D movement relative to the ground frame and sampled based on movement of the person, the accelerometer data comprising x-axis data, y-axis data and z-axis data that corresponds to the movement of the person over a period of time and storing the accelerometer data in a non-transitory memory of a computer having a processor, wherein the person's 3D movement relative to the ground frame forms a reference frame for the device;
determining tilt data indicative of the accelerometer's orientation relative to the ground frame using the accelerometer data;
generating improved accelerometer data indicative of the accelerometer's orientation relative to the reference frame for the device by filtering the accelerometer data using the tilt data;
storing the improved accelerometer data in the non-transitory memory;
generating stride data based on the improved accelerometer data, comprising:
finding by the processor a local minima and a local maxima to detect each step by the person;
finding by the processor an x-displacement along the x-axis and a y-displacement along the y-axis for each step by the person based on the improved accelerometer data for the x-axis and the y-axis and storing the x-displacement and the y-displacement in the memory;
calculating by the processor a two-dimensional (2D) movement displacement and a translation direction for each stride by the person based at least on the x-displacement and the y-displacement; and

calculating by the processor an elevation change of the person and storing the elevation change in the memory; and
estimating a 3D movement to track the location of the person, on a step by step basis based at least on the stride data.

2. The computer-implemented method as recited in claim 1 , wherein the device is a smartphone that includes an accelerometer.

3. The computer-implemented method as recited in claim 1 , wherein the accelerometer data is sampled at a frequency greater than twice a Nyquist frequency of the movement of the person.

4. The computer-implemented method as recited in claim 1 , wherein the device further includes a gyroscope adding to the x-axis data, the y-axis data and the z-axis data for the device.

5. The computer-implemented method as recited in claim 1 , wherein the local minima occurs after each heel strike and wherein the local maxima occurs after each passive position.

6. The computer-implemented method as recited in claim 1 , wherein utilizing the processor to find the local minima and the local maxima includes utilizing the processor to reduce extraneous detections by determining if there are any neighboring minima within a sample window of a first number of samples before a local minima and a second number samples after each local minima, and not counting any neighboring minima within the sample window as a local minima.

7. The computer-implemented method as recited in claim 1 , wherein the processor utilizes a neural network to classify the person's gait.

8. The computer-implemented method as recited in claim 7 , wherein the processor classifies the person's gait on a per step basis.

9. The computer-implemented method as recited in claim 1 , wherein utilizing the processor to find the x-displacement and the y-displacement is performed in a time interval between the local minima and the local maxima for each step.

10. The computer-implemented method as recited in claim 1 , further comprising utilizing the processor to determine a translation direction for each step, wherein a most frequent direction of translation is forward, and wherein the ground frame is corrected to the y-axis.

11. The computer-implemented method as recited in claim 1 , further comprising utilizing the processor to determine a translation direction for each step, wherein a most frequent direction of translation is determined by assigning each possible direction of motion based on each step to a bin for the direction of motion and the bin with the highest frequency is considered the most frequent direction of translation.

12. The computer-implemented method as recited in claim 1 , further comprising utilizing the processor to determine a translation direction for each step, and further comprising detecting a transition corresponding to an abrupt change in orientation of the device and discontinuing the step of utilizing the processor to determine a translation direction for each step until the transition has completed.

13. The computer-implemented method as recited in claim 1 , wherein calculating the 3D movement is further based on determining the person's 3D translation and rotation, and further comprising determining a location and a heading for the person based on the 3D translation and rotation.

14. The computer-implemented method as recited in claim 1 , further comprising utilizing the processor to determine a translation direction for each step, wherein even and odd steps are tracked separately to find a most frequent direction of translation for even steps and for odd steps, and a forward direction is determined as an average of the most frequent translation direction for the even steps and the odd steps.

15. The computer-implemented method as recited in claim 1 , further comprising utilizing the processor to determine a translation direction for each step, wherein a most frequent direction of translation for even steps and for odd steps is separately determined by:
assigning each direction of motion based on each step to a bin for the direction of motion; and
considering the bin with a highest frequency as the most frequent direction of translation, and wherein a forward direction is determined as an average of the most frequent translation direction for the even steps and the odd steps.

16. A computer-implemented method for classifying a person's gait, comprising:
collecting accelerometer data from a worn or person carried device that includes an accelerometer in an unknown tilted orientation relative to a ground frame as the person moves around a physical location, the accelerometer data being indicative of the person's 3D movement relative to the ground frame and sampled based on movement of the person, the accelerometer data comprising x-axis data, y-axis data and z-axis data that corresponds to the movement of the person over a period of time and storing the data in a non-transitory memory, wherein the person's 3D movement relative to the ground frame forms a reference frame for the device;
determining tilt data indicative of the accelerometer's orientation relative to the ground frame using the accelerometer data;
generating improved accelerometer data indicative of the accelerometer's orientation relative to the reference frame for the device by filtering the accelerometer data using the tilt data;
storing the improved accelerometer data in the non-transitory memory;
generating stride data based on the improved accelerometer data comprising:
detecting steps of the person within the improved accelerometer data;
inputting to a neural network a first number of values of each of the x-axis, y-axis and z-axis components of angle vectors for each step;
inputting to the neural network a second number of value

Assignees

• Trx Systems Inc

Inventors

• Kordari Kamiar
• Funk Benjamin
• Napora Jared
• Verma Ruchika
• Teolis Carole
• Young Travis

Classifications

• G06F17/10

  Complex mathematical operations {(function generation by table look-up G06F1/03; evaluation of elementary functions by calculation G06F7/544)} · CPC title

• G01P15/00

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

• G01C22/00

  Measuring distance traversed on the ground by vehicles, persons, animals or other moving solid bodies, e.g. using odometers, using pedometers · CPC title

• G01C25/005

  initial alignment, calibration or starting-up of inertial devices · CPC title

• G01C22/006 Primary

  Pedometers · CPC title