System and method for determining the activity of a mobile element

• US 9,489,600 B2
• Filed: 04/26/2010
• Issued: 11/08/2016
• Est. Priority Date: 04/24/2009
• Status: Active Grant

• Current Assignee: Commissariat a L'Energie Atomique (AES Corporation), Movea
• Original Assignee: Commissariat a L'Energie Atomique (AES Corporation), Movea
• Inventors: Jallon, Pierre
• Primary Examiner(s): Nghiem, Michael
• Assistant Examiner(s): Ngo, Peter
• Application Number:
  US13/266,094
• Publication Number:
  US 20120158351A1
• Time in Patent Office: 2,388 Days
• Field of Search: 702/141
• US Class Current: 1/1
• CPC Class Codes: A61B 2562/0219 Inertial sensors, e.g. accelerometers, gyroscopes, tilt switches
• CPC Class Codes: A61B 5/11 Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb for measuring pulse A61B5/02 ;A61B5/1038 takes precedence; motion detection to correct for motion artifacts in physiological signals A61B5/721
• CPC Class Codes: A61B 5/1101 Detecting tremor
• CPC Class Codes: A61B 5/1116 Determining posture transitions
• CPC Class Codes: G06F 18/295 Markov models or related models, e.g. semi-Markov models; Markov random fields; Networks embedding Markov models

Abstract

A system for determining activity of a mobile element, includes at least one motion sensor having a measurement axis, and which is securely connected to the mobile element. A filter for selects, for each measurement axis of the motion sensor, high frequencies above a first threshold, and processing equipment for determines a unidimensional high-frequency component (y(n)) equal to the square of the Euclidean norm of the high frequencies along the measurement axes of the motion sensor. A calculator calculates, for each state, probability density functions (Py,i) of the high-frequency component, the probability density function corresponding to each state according to a Chi-square law with a degree of freedom equal to the number of measurement axes of the motion sensor. An analyzer defines states of the mobile element, utilizing the probability density function of the high-frequency component for each state, together with the probabilities of transitions between two successive states.

Claims

1. A method for determining activity of a user having a mobile element coupled thereto, the mobile element comprising at least one motion sensor comprising an accelerometer having at least two measurement axes, the method comprising:

using the mobile element to measure movement of the user along the at least two axes to generate a corresponding signal S(n) by the motion sensor at discrete times from zero to N when the mobile element is in a state E(n) characterizing a posture of the user;

filtering the signal S(n) to obtain a first component y(n) of high-frequency portions of the signal S(n), said high frequency portions being above a first frequency threshold on each axis of the motion sensor, and a second component x(n) of low frequency portions of the signal S(n), said low frequency portions being below a second frequency threshold lower than the first frequency threshold on at least one axis of the motion sensor;

calculating by a processor a combination θ

(n) of a norm of the first component and a contribution of the second component;

determining by the processor a list of the most likely states E(k_i) characterizing the postures of the user during a period between time zero and N from a calculation of;

a probability function of obtaining θ

(n) from the signal S(n) when the mobile element is in the E(n) state;

a probability function of a transition from the E(n−

1) state to the E(n) state; and

probability functions of each state of the mobile element at time zero; and

outputting a most likely activity of the user as a sequence of most likely postures.

2. The method of claim 1, wherein the first component y(n) is a unidimensional component equal to a square of a norm of said high frequencies portions of the signal S(n) which are taken into account from k measurement axes of the motion sensor, and the probability function of said first component is a probability density function which is defined by a Chi-square law with a degree of freedom equal to k.

3. The method of claim 2, wherein the probability density function (P_y) is defined, when the mobile element is in the state i, by the following expression:

4. The method of claim 1, wherein the low-frequency portions are selected for each measurement axis of the motion sensor and the second component x(n) is of a dimension equal to the number of measurement axes of the motion sensor.

5. The method of claim 1, wherein the low frequency portions are selected for each measurement axis of the motion sensor and the second component x(n) is unidimensional.

6. The method of claim 1, wherein a probability density function (P(x(n),y(n))) for obtaining a pair of values ((x(n), y(n)) is calculated for the second component x(n) and the first component y(n), P(x(n),y(n)) being equal to a product of a probability density function of the second component x(n) and the probability density function of the first component y(n), said probability density function of the second component, when the mobile element (EM) is in the state i being defined by the following expression:

7. A system for determining activity of a user, the system including a mobile element capable of being coupled to the user, the mobile element comprising at least one motion sensor comprising an accelerometer having at least two measurement axes, said system further comprising:

a module to measure movement of the user along the at least two axes to generate a corresponding signal S(n) by the motion sensor at discrete times from zero to N when the mobile element is in a state E(n) characterizing a posture of the user;

a module for filtering the signal S(n) to obtain a first component y(n) of high-frequency portions of the signal S(n), said high frequency portions being above a first frequency threshold on each axis of the motion sensor, and a second component x(n) of low frequency portions of the signal S(n), said low frequency portions being below a second frequency threshold lower than the first frequency threshold on at least one axis of the motion sensor;

a module for determining a combination θ

(n) of a norm of the first component and a contribution of the second component;

a module for determining a list of the most likely states E(k_i,) characterizing the postures of the user during a period between time zero and N from a calculation of;

a probability function of obtaining θ

(n) from the signal S(n) when the mobile element is in the E(n) state;

a probability function of a transition from the E(n−

1) state to the E(n) state; and

probability functions of each state of the mobile element at time zero; and

an output module configured to output a most likely activity of the user as a sequence of most likely postures.

8. The system of claim 7, further comprising:

a processor configured to calculate the first component y(n) as a unidimensional high-frequency component equal to a square of a Euclidean norm of said high frequency components along each measurement axis of the motion sensor; and

a calculator configured to calculate, for each state, probability density functions of said high-frequency components, said probability density functions corresponding to each state according to a Chi-square law with a degree of freedom equal to a number k of measurement axes of the motion sensor.