System and method for detecting at least one transient phase in a steady activity of an animated being

Abstract

A system for detecting at least one transient phase in a steady activity of an animated being includes processing circuitry for determining signals representative of the motion of the animated being along at least one axis, for calculating a resultant signal representative of a statistical link between samples of signals representative of the motion belonging respectively to at least two temporally offset sliding windows over the samples, and for detecting a transient phase on the basis of the resultant signal.

Claims

1. A device, coupled to an animated being, configured for counting transitions along at least one pathway, the device comprising: at least one motion sensor; processing circuitry configured to receive, from the at least one motion sensor, signals representative of motion of said animated being along the at least one pathway, wherein said signals representative of the motion are determined along at least two pathways, each pathway having at least one component; the processing circuitry further configured to determine, based on an electronic assessment of the determined signals, a resultant signal representative of a statistical link between samples of said signals representative of said motion belonging respectively to at least two temporally offset sliding windows over said samples, wherein a steady activity includes an invariance of the statistical link of the samples of said signals representative of said motion, wherein the processing circuitry is further configured to merge the at least one component on each pathway of said resultant signal; the processing circuitry further configured to detect at least one transient phase of a plurality of transient phases, based on an electronic assessment of said resultant signal, each transient phase corresponding to a transition from a first pathway to a second pathway during which there is a modification of the statistical link of the samples of said signals representative of said motion, wherein the second pathway is in a different direction than the first pathway; the processing circuitry further configured to generate a count of a number of detected at least one transient phases, wherein the generated count indicates a number of pathways; and the processing circuitry further configured to cause the generated count to be sent in an output signal.

2. The device as claimed in claim 1, further comprising: the processing circuitry further configured to filter said signals representative of said motion prior to the calculating.

3. The device as claimed in claim 1, wherein said sliding windows partially overlap.

4. The device as claimed in claim 1, wherein said sliding windows are of the same size.

5. The device as claimed in claim 1, wherein the sliding windows, temporally ordered and respectively indexed from 1 to F, are temporally spaced apart so as to observe the following relationship:
T.sub.1+D.sub.FS in which: T.sub.1 represents a duration of a first window indexed 1, D.sub.F represents an offset between the first window indexed 1 and a last window indexed F, and S corresponds to a minimum duration of steady activity between two successive transient phases.

6. The device as claimed in 1, wherein the processing circuitry is further configured to calculate a sum of components along each pathway of said resultant signal.

7. The device as claimed in claim 6, wherein the processing circuitry is further configured to: detect a first transient phase in steady activity when the resultant signal along a single pathway or a merged signal output becomes greater than a first threshold; and detect another transient phase in the steady activity when the resultant signal along the single pathway or the merged signal output, since a prior detection of a transient phase, has been lower than a second threshold and then is higher than said first threshold, wherein said second threshold being lower than said first threshold.

8. The device as claimed in claim 7 wherein the processing circuitry is further configured to establish a difference between two instants of said signals representative of the motion, and configured to determine a maximum energy component of differentiated signals, wherein the processing circuitry is further configured to detect another transient phase in the steady activity when a sign of said maximum component, when the merged signal is above said first threshold, is opposite to a sign of said maximum component in a preceding detection.

9. The device as claimed in claim 1, wherein the processing circuitry is further configured to determine timing of the transient or steady phases.

10. The device as claimed in claim 1 wherein the processing circuitry is further configured to detect half-turns of a person, between two crossings of a straight line in opposite directions, or to detect changes of direction of a person on a course comprising a series of straight lines.

11. The device as claimed in claim 10, wherein the number of pathways indicates to and return lengths of a swimmer in a swimming pool.

12. The device as claimed in claim 1, wherein the processing circuitry is disposed within a sealed box coupled to a portion of a body of a swimmer and is further configured for determining signals representative of motion of said swimmer, said steady activity being swimming of a swimming pool length and the transient phase being a half-turn.

13. The device as claimed in claim 1, wherein the statistical link comprises a covariance function.

14. A method for use in a device, coupled to an animated being, for counting transitions along at least one pathway, the method comprising: acquiring, from at least one motion sensor, signals representative of motion of said animated being determined along the at least one pathway, wherein said signals representative of the motion are determined along at least two pathways, each pathway having at least one component; determining, based on an electronic assessment of the acquired signals, a resulting signal representative of a statistical link between samples of said signals representative of said motion belonging respectively to at least two temporally offset sliding windows over said samples, wherein a steady activity includes an invariance of the statistical link of the samples of said signals representative of said motion, wherein the at least one component is merged on each pathway of said resultant signal; detecting at least one transient phase of a plurality of transient phases, based on an electronic assessment of said resultant signal, each transient phase corresponding to a transition from a first pathway to a second pathway during which there is a modification of the statistical link of the samples of said signals representative of said motion, wherein the second pathway is in a different direction than the first pathway; generating a count of a number of detected at least one transient phases, wherein the generated count indicates a number a number of pathways; and causing the generated count to be sent in an output signal.

15. A device, coupled to swimmer, configured for counting swimming pool lengths, the device comprising: at least one motion sensor; processing circuitry configured to receive, from the at least one motion sensor, signals representative of motion of said swimmer being along a swimming pool length, wherein said signals representative of the motion are determined along at least two pathways, each pathway having at least one component; the processing circuitry further configured to determine, based on an electronic assessment of the determined signals, a resultant signal representative of a statistical link between samples of said signals representative of said motion belonging respectively to at least two temporally offset sliding windows over said samples, wherein a steady activity includes an invariance of the statistical link of the samples of said signals representative of said motion, wherein the processing circuitry is further configured to merge the at least one component on each pathway of said resultant signal; the processing circuitry further configured to detect at least one transient phase of a plurality of transient phases, based on an electronic assessment of said resultant signal, each transient phase corresponding to a transition from a first swimming pool length to a second swimming pool length during which there is a modification of the statistical link of the samples of said signals representative of said motion, wherein the second swimming pool length is in a reverse direction of the first swimming pool length; the processing circuitry further configured to generate a count of a number of detected at least one transient phases, wherein the generated count indicates a number of swimming pool lengths; and the processing circuitry further configured to cause the generated count to be sent in an output signal.

16. The device as claimed in claim 15, wherein the processing circuitry is further configured to: detect a first transient phase in steady activity when the resultant signal along a single swimming pool length or a merged signal output becomes greater than a first threshold; and detect another transient phase in the steady activity when the resultant signal along the swimming pool length or the merged signal output, since a prior detection of a transient phase, has been lower than a second threshold and then is higher than said first threshold, wherein said second threshold being lower than said first threshold.

17. The device as claimed in claim 16, wherein the processing circuitry is further configured to establish a difference between two instants of said signals representative of the motion, and configured to determine a maximum energy component of differentiated signals, wherein the processing circuitry is further configured to detect another transient phase in the steady activity when a sign of said maximum component, when the merged signal is above said first threshold, is opposite to a sign of said maximum component in a preceding detection.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The invention will be better understood on studying a few embodiments described as non-limiting examples and illustrated by the attached drawings in which:

(2) FIGS. 1 to 5 schematically illustrate embodiments of a system, according to one aspect of the invention; and

(3) FIGS. 6 to 13 illustrate signals corresponding to the processing operations implemented by the system of FIG. 4, according to one embodiment of the invention.

DETAILED DESCRIPTION OF THE INVENTION

(4) In FIG. 1, the system for detecting at least one transient phase in a steady activity of an animated being, according to one embodiment of the invention, comprises a module DETER for determining signals representative of the motion of said animated being along at least one axis, a module CALC for calculating a resultant signal representative of a statistical link between samples of said signals representative of said motion belonging respectively to at least two temporally offset sliding windows, and a module DETEC for detecting a transient phase on the basis of said resultant signal.

(5) An optional module FILT for filtering said signals representative of said motion upstream of said calculation means, as represented in FIG. 2.

(6) For example, the sliding windows can partially overlap, and/or be of the same size.

(7) Furthermore, the temporally ordered sliding windows respectively indexed from 1 to F are temporally spaced apart so as to observe the following relationship:
T.sub.1+D.sub.FS
in which: T.sub.1 represents the duration of the first window indexed 1, D.sub.F represents the offset between the first window indexed 1 and the last window indexed F, and S corresponds to a minimum duration of the steady activity between two successive transient phases.

(8) Furthermore, as illustrated in FIG. 3, the system can comprise a module FUS for merging the components on each pathway of said resultant signal, for example adapted to calculate the sum of the components along each pathway of said resultant signal.

(9) Furthermore, the detection module DETEC can be adapted to: detect a first transient phase in the steady activity when the resultant signal along a single pathway or the merged signal output from the merging means FUS becomes greater than a first threshold Seuil_1; and detect another transient phase in the steady activity when the resultant signal along a single pathway or the merged signal output from the merging means FUS, since the last detection of a transient phase, has been lower than a second threshold Seuil_2 than is higher than said first threshold Seuil_1;
said second threshold Seuil_2 being lower than said first threshold Seuil_1.

(10) Said thresholds Seuil_1 and Seuil_2 can depend on the measurement range of the sensor or sensors used to acquire said signals representative of the motion, and/or are set by learning from a base of motion signals recorded by the sensor or sensors.

(11) Furthermore, the system can comprise a module CPT for counting the number of transient phases, as illustrated in FIG. 4.

(12) The detection module DETEC can be adapted to detect half-turns of a person, between two crossings of a straight line in opposite directions.

(13) For example, the system can comprise a sealed box comprising the means DETER for determining signals representative of the motion of a swimmer and fixing means for securely linking the box to a part of the body of the swimmer, said steady activity being the swimming of a swimming pool length and the transient phase being a half-turn. The counting module CPT can be adapted to count the go and return lengths of a swimmer in a swimming pool. Also, the counting module can deduce therefrom the number of lengths travelled, therefore the distance that has been swum. It can also provide the instants corresponding to the half-turns detected. It will be possible to deduce therefrom an average swimming speed, for example, over one or more lengths.

(14) The calculation of the statistical link can, for example, comprise a covariance function.

(15) Furthermore, as illustrated in FIG. 5, the system can comprise a differentiation module DIFF adapted to establish a difference between two instants t and t+t (with t<S) of the signals representative of the motion, and a module DET_CEM for determining the maximum energy component of the differentiated signals. The detection module DETEC is then adapted to detect another transient phase in the steady activity when, in addition, the sign of said maximum component, when the merged signal is above the first threshold Seuil_1, is the opposite of the sign of the maximum component upon the preceding detection. The energy of a signal corresponds to the sum of its samples squared. The number of false detections is thus limited.

(16) The following example is described in a non-limiting manner.

(17) The embodiment calculation module CALC is adapted to supply a signal resulting from a function which captures the statistical relationship between the parts of the signals obtained from at least two temporally offset sliding windows, from at least one signal representative of the motion covered by the sliding windows. The resultant signal is characterized by peaks at the time of the transient phases and transitions through values close to zero during the steady phases. It is then possible, by setting decision thresholds, to detect and count the transient phases automatically.

(18) For example, it is possible to detect the half-turns in a series of lengths executed by a swimmer or a sprinter. For example, the signals representative of the motion are obtained from a triaxial magnetometer worn by the sports person. Also, the determination module DETER comprises the triaxial magnetometer.

(19) The triaxial magnetometer of the determination module DETER fixed to the body of the swimmer or of the sprinter (head, back, ankle or wrist) supplies a sample signal M.sub.i(t), i{x, y, z}, t which changes as a function of the go or return direction of the sports person. The signal Mx, My, Mz on the three axes x, y and z of the magnetometer is represented in FIG. 6, corresponds to the measurement of sixteen steady phases intercepted by fifteen transient phases.

(20) The filtering module FILT makes it possible to attenuate the spurious motions that are not characteristic of the transient phases and which can disturb their detection. These motions appear during the steady activity and are due, in the present example, to the strokes of the swimmer or to the strides of the sprinter. In this case, the filtering can be of bypass type (average filter) in order to eliminate the spurious motions (high frequencies) without affecting the useful low-frequency information linked to the half-turns as illustrated in FIG. 7 for each axis x, y, z:

(21) ${\mathrm{MF}}_i\left(t\right)=\frac{1}{L}\underset{n=t-L+1}{\overset{t}{.Math.}}{M}_i\left(n\right),i\left\{x,y,z\right\},L$

(22) The size L of the sliding window which is for calculating the average is preferably configured such that E<L<S, in which S is the minimum duration of the steady activity between two transient phases and E represents the maximum deviation between two spurious motions.

(23) The filtering step is optional in the processing chain, but its application can make it possible to improve the detection of the transient phases. It is possible to replace the average with other types of filter that make it possible to attenuate or eliminate the components of the signal that correspond to the motions which do not convey information on the transient phase.

(24) The calculation module CALC performs a windowing, as illustrated in FIG. 8, which includes defining F sliding windows (F*\{1}) which cover the components of the motion signal to be analyzed. The F temporally ordered windows can be positioned in such a way as to observe the following relationship:
T.sub.1+D.sub.FS
in which: T.sub.1 represents the duration of the first window indexed 1, D.sub.F represents the offset between the last sample of the first window indexed 1 and the first sample of the last window indexed F, and S still represents the minimum duration of the steady activity between two transient phases.

(25) If the first sliding window indexed 1, of size T.sub.1 considered at the instant t, picks up the filtered signal MF as follows:
MF1.sub.i.sup.(t)(k)=MF.sub.i(k+tT.sub.1), k[1;T.sub.1], i{x,y,z}

(26) Then the F1 other sliding windows indexed j of size T.sub.j pick up the filtered signal MF at the instants t+D.sub.j by observing the following relationship:
MFj.sup.(t+D.sup.j.sup.)(k)=MF.sub.i(k+t+D.sub.jT.sub.j), k[1;T.sub.j], i{x,y,z}
in which D.sub.j represents the offset between the last sample of the first window of index 1 and the window of index j.

(27) By positing D.sub.1=0, it is possible to define, generically, the signals picked up by the F sliding windows at a given instant t
MFj.sub.i.sup.(t+D.sup.j.sup.)(k)=MF.sub.i(k+t+D.sub.jT.sub.j), k[1;T.sub.j], i{x,y,z}, j[1;F]

(28) Before estimating the relationship or the statistical link between the signals obtained from the F sliding windows, the situation is reduced to the case in which the signals are all of the same size.

(29) To reduce the situation to the case of signals comprising N samples, it is possible to remove (respectively add), by decimation (respectively by interpolation) T.sub.jN (respectively NT.sub.j) samples evenly spaced apart (respectively intercalated) within the signal from the window of index j and of initial size T.sub.j>N (respectively T.sub.j<N).

(30) The calculation of the statistical link includes calculating a signal MFC resulting from a function f which models, for each measurement axis, the statistical link that exists between the signals of size N from the F sliding windows defined at the instant t as follows:
MFC.sub.i(t)=f(MF1.sub.i.sup.(t), . . . , MFj.sub.i.sup.(t+D.sup.j.sup.), . . . , MFF.sub.i.sup.(t+D.sup.F.sup.)), i{x,y,z}

(31) The function f which picks up the statistical link between the signals from the F sliding windows can take the following form:

(32) $f\left(\mathrm{MF}{1}_i^{\left(t\right)},\mathrm{.Math.},{\mathrm{MFj}}_i^{\left(t+D_j\right)},\mathrm{.Math.},{\mathrm{MFF}}_i^{\left(t+D_F\right)}\right)=\underset{n=1}{\overset{N}{.Math.}}\underset{j=1}{\overset{F}{.Math.}}{\left({\mathrm{MFj}}_i^{\left(t+D_j\right)}\left(n\right)-j_i^{\left(t+D_j\right)}\right)}^{p_j}$$\mathrm{with}{j}_i^{\left(t+D_j\right)}=\frac{1}{N}\underset{k=1}{\overset{N}{.Math.}}{\mathrm{MFj}}_i^{\left(t+D_j\right)}\left(k\right)\mathrm{and}{p}_j$

(33) FIG. 9 represents a few examples of signals MFC (with one component) resulting from the function f when the number of sliding windows F is equal to two, for different pairs of values (p.sub.1, p.sub.2).

(34) The function f picks up the statistical link between the signals from F=2 sliding windows which cover the filtered signal MF with one measurement axis (a). The resulting signal is shown for different values of p1 and p2: p1=p2=1 (b), p1=1 and p2=2 (c), p1=p2=2 (d). In all the cases, the 15 peaks indicate the 15 transient phases.

(35) FIG. 10 represents a signal MFC with 3 components obtained in the case where the number of sliding windows F=2 and for which the exponents p1=p2=1. It should be noted that, for which precise case, the function f corresponds to the covariance function.

(36) The merging module FUS makes it possible to condense the information by transforming the multi-pathway or multi-component signal, in this case three components, a MFC.sub.i, i{x,y,z} into a signal with a single component MFC with:

(37) $\mathrm{MFC}\left(t\right)=\underset{i}{.Math.}{\mathrm{MFC}}_i\left(t\right)$
as illustrated in FIG. 11.

(38) Before the detection step, it appears that the merged signal exhibits the characteristics desirable to facilitate the detection of the transient phases, namely: a succession of peaks at the half-turns intercepted by 0 bands during the steady activity (swimming or sprinting phases).

(39) The detection module DETEC detects a first transient phase when the amplitude of the merged signal MFC becomes greater than a first threshold value seuil1. Following a turn numbered j detected at the instant T, a turn numbered j+1 will then be detected at the instant T+1 provided that MFC(T+1)>seuil1 and that the amplitude of the signal has dropped back below a second threshold value seuil2 between the instants T and T+1. In practice seuil1>seuil2 is illustrated in FIG. 12. The threshold values are, for example, defined by learning on an annotated database.

(40) The black stars in FIG. 12 represent the real instants of the half-turns of the sprinter or of the swimmer whereas the squares indicate the instants of the half-turns detected automatically by the system. The horizontal lines respectively symbolize the thresholds seuil1 and seuil2.

(41) The efficiency with which transient phases are detected by the system can be enhanced by adding an additional constraint at the decision-taking time, for example by involving the sign of the main component (the highest energy component) of the filtered signal, differentiated between the instants t and tt, as illustrated in FIG. 13:
MFD.sub.i(t)=MF.sub.i(t)MF.sub.i(tt), t, i{x,y,z}
t being of the same order of magnitude as the size L of the window used for calculating the average.

(42) In the figure, the filtered and differentiated signal with three components has, for the maximum component, the component z, which changes sign on each half-turn.

(43) In addition to the conditions on the amplitude of the signal MFC described previously, a turn can be detected if the sign of the maximum component at the moment when the first threshold seuil1 is exceeded by MFC is different from the sign of this same component during the preceding turn. This additional constraint makes it possible to reduce the number of false detections of transient phases but, on the other hand, does make it necessary to differentiate the signal MF which otherwise is not necessary in the proposed processing chain. It should be noted that the use of this sign, even if it improves the results, is totally optional and does not in any way modify the calculations of the processing chain.

(44) Those skilled in the art will recognize that the present invention has many applications, may be implemented in various manners and, as such is not to be limited by the foregoing embodiments and examples. Any number of the features of the different embodiments described herein may be combined into a single embodiment, the locations of particular elements can be altered and alternate embodiments having fewer than or more than all of the features herein described are possible. Functionality may also be, in whole or in part, distributed among multiple components, in manners now known or to become known.

(45) It will be appreciated by those skilled in the art that changes could be made to the embodiments described above without departing from the broad inventive concept thereof. It is understood, therefore, that this invention is not limited to the particular embodiments disclosed, but it is intended to cover modifications within the spirit and scope of the present invention. While there has been shown and described fundamental features of the invention as applied to being exemplary embodiments thereof, it will be understood that omissions and substitutions and changes in the form and details of the disclosed invention may be made by those skilled in the art without departing from the spirit of the invention. Moreover, the scope of the present invention covers conventionally known, future developed variations and modifications to the components described herein as would be understood by those skilled in the art.