DE102018222584A1 - Method for determining steps of a living being, in particular a person - Google Patents

Abstract

The invention relates to a method for determining a number of movements of a living being, comprising the steps of providing at least one measurement signal for measured movements of a person using at least one inertial sensor, providing a plurality of different types of movement of a person, each type of movement having a different periodic sequence corresponds to movements, and each movement type is assigned a positive integer n, at least one of the assigned numbers being greater than 1, - determining a current movement type of the person on the basis of the measurement signal, - providing at least one evaluation event for the determined current movement type, wherein the at least one evaluation event corresponds to the nth occurrence of a movement in the periodic sequence of movements of a person, - providing an event frequency for the occurrence of the at least one evaluation event in the measurement signal, - providing ellen of at least one evaluation rule for evaluating the measurement signal, a determination event being determined on the basis of the evaluation rule when an evaluation event corresponding to the event frequency occurs - applying the at least one evaluation rule to the measurement signal, - determining a number of determination events based on the applied evaluation rule, - determining the number of movements based on the number n and the determined number of detection events.

Description

Technical field

The invention relates to a method for determining a number of movements of a living being, in particular a person.

The invention further relates to a system for determining steps of a living being, in particular a person.

State of the art

Although the present invention is generally applicable to any method of determining a number of movements of a living being, the present invention is described with respect to methods of determining a person's steps based on measured acceleration values.

Known methods for determining a number of steps of a person, also called pedometer for short, use an acceleration sensor or a gyroscope to generally determine a person's movement, more precisely to determine from the measured data whether the person has taken a step or not. Depending on the application, various algorithms and methods that analyze the sensor signals have become known for evaluating the sensor data. Based on the analysis, it is then determined whether there is a step or not. Depending on where these sensors are located on or with a person, i.e. in which environment they are used, the accuracy of the determination of steps varies. For example, an acceleration sensor provides more accurate data for detecting steps when attached to the head compared to attaching it to an arm of a person.

Disclosure of the invention

• - Bereitstellen zumindest eines Messsignals für gemessene Bewegungen einer Person anhand zumindest eines Inertialsensors,
• - Bereitstellen einer Mehrzahl von unterschiedlichen Bewegungsarten einer Person, wobei jede Bewegungsart zu einer unterschiedlichen periodischen Abfolge von Bewegungen korrespondiert, und wobei jeder Bewegungsart eine positive ganze Zahl n zugeordnet wird, wobei zumindest eine der zuordneten Zahlen größer als 1 ist,
• - Ermitteln einer aktuellen Bewegungsart der Person anhand des Messsignals,
• - Bereitstellen von zumindest einem Auswerteereignis für die ermittelte aktuelle Bewegungsart, wobei das zumindest eine Auswerteereignis zu dem n-ten Auftreten einer Bewegung in der periodischen Abfolge von Bewegungen einer Person korrespondiert,
• - Bereitstellen einer Ereignisfrequenz für das Auftreten des zumindest einen Auswerteereignisses im Messsignal,
• - Bereitstellen von zumindest einer Auswerteregel zum Auswerten des Messsignals, wobei ein Ermittlungsereignis anhand der Auswerteregel dann festgestellt wird, wenn ein Auswerteereignis entsprechend der Ereignisfrequenz auftritt
• - Anwenden der zumindest einen Auswerteregel auf das Messsignal,
• - Ermitteln einer Anzahl von Ermittlungsereignissen anhand der angewandten Auswerteregel,
• - Ermitteln der Anzahl der Bewegungen basierend auf der Zahl n und der ermittelten Anzahl der Ermittlungsereignisse.

In one embodiment, the invention provides a method for determining periodic movements of a living being, in particular a person, comprising the steps

• Provision of at least one measurement signal for measured movements of a person using at least one inertial sensor,
• Providing a plurality of different types of movement of a person, each type of movement corresponding to a different periodic sequence of movements, and each type of movement being assigned a positive integer n, at least one of the assigned numbers being greater than 1,
• Determining a current type of movement of the person on the basis of the measurement signal,
• Provision of at least one evaluation event for the determined current type of movement, the at least one evaluation event corresponding to the nth occurrence of a movement in the periodic sequence of movements of a person,
• Providing an event frequency for the occurrence of the at least one evaluation event in the measurement signal,
• - Providing at least one evaluation rule for evaluating the measurement signal, a determination event based on the evaluation rule is then determined when an evaluation event occurs according to the event frequency
• Applying the at least one evaluation rule to the measurement signal,
• - determining a number of investigative events based on the evaluation rule applied,
• - Determining the number of movements based on the number n and the determined number of detection events.

• - Bereitstellen einer Mehrzahl von unterschiedlichen Bewegungsarten einer Person, wobei jede Bewegungsart zu einer unterschiedlichen periodischen Abfolge von Bewegungen korrespondiert, und wobei jeder Bewegungsart eine positive ganze Zahl n zugeordnet wird, wobei zumindest eine der zuordneten Zahlen größer als 1 ist,
• - Ermitteln einer aktuellen Bewegungsart der Person anhand des Messsignals,
• - Bereitstellen von zumindest einem Auswerteereignis für die ermittelte aktuelle Bewegungsart, wobei das zumindest eine Auswerteereignis zu dem n-ten Auftreten einer Bewegung in der periodischen Abfolge von Bewegungen einer Person korrespondiert,
• - Bereitstellen einer Ereignisfrequenz für das Auftreten des zumindest einen Auswerteereignisses im Messsignal,
• - Bereitstellen von zumindest einer Auswerteregel zum Auswerten des Messsignals, wobei ein Ermittlungsereignis anhand der Auswerteregel dann festgestellt wird, wenn ein Auswerteereignis entsprechend der Ereignisfrequenz auftritt
• - Anwenden der zumindest einen Auswerteregel auf das Messsignal,
• - Ermitteln einer Anzahl von Ermittlungsereignissen anhand der angewandten Auswerteregel,
• - Ermitteln der Anzahl der Bewegungen basierend auf der Zahl n und der ermittelten Anzahl der Ermittlungsereignisse.

In a further embodiment, the invention provides a system for determining periodic movements of a living being, in particular a person, comprising at least one inertial sensor for measuring movements of a person and for providing at least one measurement signal for the measured movements, a determination device which is connected to the inertial sensor and which is designed to carry out the following steps:

• Providing a plurality of different types of movement of a person, each type of movement corresponding to a different periodic sequence of movements, and each type of movement being assigned a positive integer n, at least one of the assigned numbers being greater than 1,
• Determining a current type of movement of the person on the basis of the measurement signal,
• Provision of at least one evaluation event for the determined current type of movement, the at least one evaluation event corresponding to the nth occurrence of a movement in the periodic sequence of movements of a person,
• Providing an event frequency for the occurrence of the at least one evaluation event in the measurement signal,
• - Providing at least one evaluation rule for evaluating the measurement signal, a determination event being determined on the basis of the evaluation rule when an evaluation event corresponding to the event frequency occurs
• Applying the at least one evaluation rule to the measurement signal,
• - determining a number of investigative events based on the evaluation rule applied,
• - Determining the number of movements based on the number n and the determined number of detection events.

One of the advantages achieved in this way is that the accuracy in the determination of periodic movements of a living being, in particular the number of steps of a person, is increased, in particular the number of false negative or false positive steps which falsify the number of steps is reduced . Another advantage is that a simple and inexpensive implementation or a simple feasibility of the method is made possible. Another advantage is the higher flexibility, since any detection method can be used to determine periodic movements. In other words, it is particularly avoided in this way that when the inertial sensor is arranged on the wrist, the arm movement overlaps or falsifies the leg movement of the person, which leads to a large number of undetected steps - false negative steps.

Further features, advantages and further embodiments of the invention are described below or become apparent thereby.

According to an advantageous development, n = 2 or n = 4 is selected for the number n. This enables the number of steps of the person to be determined simply and precisely. In particular, when n = 4, periodic movements of four-footed animals or the like can also be reliably recognized and their steps counted.

According to a further advantageous development, the set of at least two predetermined types of movement includes at least two of the following movement states: “walking”, “walking”, “Nordic walking”, “running”, “jogging”, “hopping” and / or “sprinting” ". The advantage of this is an exact determination of the current state of movement of a person. Other movement states are generally conceivable for living beings, for example different gaits of horses or the like.

According to a further advantageous development, the measurement signal is analyzed to determine a measurement signal property, in particular in which a frequency analysis of the measurement signal is carried out, and the current type of movement is determined from the plurality of movement types provided on the basis of the determined measurement signal property. Measurement signal properties include, for example, statistical variables such as mean strength / amplitude, standard deviation, frequency / period, and their respective standard deviation and the like. One of the advantages achieved with this is that the accuracy in determining the current state of motion is further improved.

According to a further advantageous development, the event frequency is provided on the basis of the current type of movement determined. The advantage of this is a quick and at the same time reliable determination of the periodic movement of a living being.

According to a further advantageous development, the at least one evaluation rule takes into account a deviation interval of the event frequency. The accuracy when recognizing periodic movements of a living being can thus be further improved by, within certain limits, taking deviations from the sampling rate or its frequency into account as a detection event.

According to a further advantageous development, the evaluation rule is provided in such a way that peaks of a sum signal, in particular an amount of an acceleration vector in at least two spatial directions, are evaluated in order to determine a determination event. The accuracy or robustness is thus further improved, in particular the effect of offsets on provided acceleration signals being reduced.

According to a further advantageous development, an inertial sensor is provided as an acceleration sensor. This provides a simple, inexpensive inertial sensor.

According to a further advantageous development, a statistical evaluation of provided acceleration values takes place in order to provide the at least two predetermined types of movement. Different types of movement can thus be provided in a simple and reliable manner.

According to a further advantageous development, the statistical evaluation takes place on the basis of a calculation of mean values and variances of the acceleration values provided, in particular wherein the acceleration values provided are smoothed exponentially. This enables statistical evaluation to be carried out quickly and at the same time reliably.

According to a further advantageous development, the at least one inertial sensor is designed as an acceleration and / or rotation rate sensor, in particular in the form of a multi-axis sensor. A rotation rate sensor can thus be provided in a simple and inexpensive manner, in particular in the case of an inertial sensor in the form of an acceleration sensor.

It goes without saying that the features mentioned above and those yet to be explained below can be used not only in the combination specified in each case, but also in other combinations or on their own without departing from the scope of the present invention.

Preferred embodiments and embodiments of the invention are shown in the drawings and are explained in more detail in the following description, the same reference numerals referring to the same or similar or functionally identical components or elements.

Figure list

• 1 ein System zum Ermitteln von Schritten einer Person gemäß einer Ausführungsform der vorliegenden Erfindung;
• 2 Teile eines Verfahrens gemäß einer Ausführungsform der vorliegenden Erfindung;
• 3 Teile eines Verfahrens gemäß einer Ausführungsform der vorliegenden Erfindung;
• 4 zeigt ein Schaubild eines Gyroskopsignals und eines Beschleunigungssignals über die Zeit für eine Bewegungsart gemäß einer Ausführungsform der vorliegenden Erfindung; und
• 5 Schritte eines Verfahrens gemäß einer Ausführungsform der vorliegenden Erfindung.

It shows

• 1 a system for determining steps of a person according to an embodiment of the present invention;
• 2nd Parts of a method according to an embodiment of the present invention;
• 3rd Parts of a method according to an embodiment of the present invention;
• 4th FIG. 4 shows a graph of a gyroscope signal and an acceleration signal over time for one type of movement according to an embodiment of the present invention; FIG. and
• 5 Steps of a method according to an embodiment of the present invention.

Embodiments of the invention

1 10 shows a system for determining steps of a person according to an embodiment of the present invention.

Weiterhin kann auf die berechneten Beschleunigungswerte ein digitaler Butterworth-Filter zweiter Ordnung mit einer Abschneidefrequenz zwischen 4-7 Hz bei einer Abtastrate von 50 Hz angewendet werden, um Hochfrequenzrauschen in der Stärke der berechneten Beschleunigungswerte zu entfernen. Die Abschneidefrequenz ergibt sich hierbei beispielsweise aus der Schrittfrequenz: 4 Hz entspricht 4 oder 2 Schritten pro Sekunde, je nachdem ob jeder oder nur jeder zweite Schritt im Beschleunigungssignal erkannt wird. In einer weiteren Ausführungsform kann die Abschneidefrequenz auch der zu erwartenden Schrittfrequenz angepasst werden beispielsweise, um das Verfahren auch bei höheren Schrittfrequenzen wie bei Tieren anwenden zu können.In detail is in 1 a system 1 shown. The system 1 includes an acceleration sensor 2nd , the acceleration values as a function of three spatial directions for a preprocessing unit 3rd provides. The preprocessing unit 3rd calculates the strength of the acceleration signal using formula I: $$ac{c}_{rms}=\sqrt{ac{c}_x^{\mathrm{2nd}}+ac{c}_y^{\mathrm{2nd}}+ac{c}_{\mathrm{e.g.}}^{\mathrm{2nd}}}$$

Furthermore, a digital second order Butterworth filter with a cut-off frequency between 4-7 Hz at a sampling rate of 50 Hz can be applied to the calculated acceleration values in order to remove high frequency noise in the strength of the calculated acceleration values. The cut-off frequency results, for example, from the step frequency: 4 Hz corresponds to 4 or 2 steps per second, depending on whether every or only every second step is recognized in the acceleration signal. In a further embodiment, the cut-off frequency can also be adapted to the step frequency to be expected, for example, in order to be able to use the method at higher step frequencies as in animals.

The preprocessing unit 3rd then transmits the preprocessed signal with the calculated acceleration values to the step determination device 4th . The step determination device 4th comprises a movement type determination unit 4a which determines the type of movement of the person based on the preprocessed signal based on the provided types of movement. On the basis of the movement type determined in this way, a time period between two successive steps is then determined, a step counting unit 4b configured with this period of time and the preprocessed acceleration signal transmitted to it. The pedometer unit 4b now counts steps according to the configured time, whereby steps that occur outside the configured time may not be counted. The number of steps counted is then sent to a multiplier 4c that multiplies the number by a factor that takes into account the steps that are not counted. Then the determined number of steps 6 from the step determination device 4th issued or provided.

2nd shows parts of a method according to an embodiment of the present invention.

berechnet wurden, wobei α einen Glättungsfaktor im Intervall [0,1] darstellt.In 2nd is the mean of the acceleration 7 over the variance of acceleration 8th plotted for a large number of provided acceleration values of different data sets, the mean and variance using exponential smoothing filters according to the recursive formula $$y\lfloor n\rfloor =\left(1-\alpha \right)y\left[n-1\right]+\alpha \times \left[n\right]$$

were calculated, where α represents a smoothing factor in the interval [0.1].

A classification is then carried out on the basis of provided acceleration data. With reference to the present 2nd will be a linear discriminant function for the Classification selected for its simplicity and the resulting low memory consumption. Other classification functions, for example non-linear function or the like, for example a support vector machine or for example artificial neural networks, are also possible.

3rd shows parts of a method according to an embodiment of the present invention.

In detail is in 3rd an acceleration 11 over time 10 applied. A lower envelope curve is shown in each case 12a and an upper envelope 12b for the strength of the acceleration 15 . Between the upper and lower envelope 12a , 12b is the average strength of the acceleration based on the broken line 13 shown. steps 14 are detected when the upper envelope 12b is touched after the lower envelope 12a was touched by the measured or the calculated acceleration value. In particular, it can be taken into account here that the strength of the acceleration must be above a certain threshold value, as well as the time difference between two successive steps, which must also be above a certain threshold value. In addition to the in 3rd described method is also a zero crossing / average method for determining steps or any other peak detection method for determining steps.

4th FIG. 12 shows a graph of a gyroscope signal and an acceleration signal over time for one type of movement according to an embodiment of the present invention.

In detail is in 4th above a gyroscope signal 11a in rad / s in the z direction, i.e. in the vertical direction over time 10 shown, an acceleration signal in the middle 11 in the x, y, and z directions and in 4th below the amount 11 of the acceleration vector according to formula I, each in m / s ² when a person walks and moves his arms. It can be seen that the signal 11a the gyroscope provides a good, clear signal for recognizing different steps; the turning points are essentially unique. In contrast to this is the signal 11 based on the magnitude of the acceleration vector not very clear and of lower quality than that of the gyroscope. By comparing the signals of the 4th above and the 4th The respective measured steps can be recognized or assigned below. Steps in the diagram of the 4th below, which are marked with the arrows, are based on the signal 11 the amount of the acceleration vector can only be recognized or identified at all with high uncertainty, whereas the respective alternative steps are based on the peaks in the signal 11 of the amount of the acceleration vector - marked in the 4th below with circles - are clearly recognizable and identifiable. In this respect, the total number of steps of the person can be determined by only recognizing and counting every second step and the number thus determined with the factor 2nd is multiplied. So despite a signal 11 poorer quality with high accuracy the total number of steps covered can be determined.

5 shows steps of a method according to an embodiment of the present invention.

In 5 a method of determining a number of movements of a living being is shown. The process comprises the following steps.

In a first step S1 at least one measurement signal is provided for measured movements of a person using at least one inertial sensor.

In a further step S2 a plurality of different types of movement of a person is provided, each type of movement corresponding to a different periodic sequence of movements, and each type of movement being assigned a different positive integer n.

In a further step S3 a current movement type of the person is determined on the basis of the measurement signal.

In a further step S4 at least one evaluation event is provided for the determined current type of movement, the at least one evaluation event corresponding to the nth occurrence of a movement in the periodic sequence of movements of a person.

In a further step S5 an event frequency is provided for the occurrence of the at least one evaluation event in the measurement signal.

In a further step S6 at least one evaluation rule is provided for evaluating the measurement signal, a determination event being determined on the basis of the evaluation rule when an evaluation event corresponding to the event frequency occurs.

In a further step S7 the at least one evaluation rule is applied to the measurement signal.

In a further step S8 a number of ascertainment events are ascertained on the basis of the applied evaluation rule.

In a further step S9 the number of movements is determined based on the number n and the determined number of detection events.

• • Einfache Implementierung.
• • Kostengünstige Implementierung.
• • Höhere Genauigkeit.
• • Verbesserte Robustheit.

In summary, at least one of the embodiments of the invention has at least one of the following advantages:

• • Easy implementation.
• • Inexpensive implementation.
• • Higher accuracy.
• • Improved robustness.

Although the present invention has been described on the basis of preferred exemplary embodiments, it is not restricted to these but can be modified in a variety of ways.

Claims (12)

A method of determining a number of movements of a living being, comprising the steps Providing (S1) at least one measurement signal for measured movements of a person using at least one inertial sensor, Providing (S2) a plurality of different types of movement of a person, each type of movement corresponding to a different periodic sequence of movements, and each type of movement being assigned a positive integer n, at least one of the assigned numbers being greater than 1, Determining (S3) a current type of movement of the person on the basis of the measurement signal, Providing (S4) at least one evaluation event for the determined current type of movement, the at least one evaluation event corresponding to the nth occurrence of a movement in the periodic sequence of movements of a person, Providing (S5) an event frequency for the occurrence of the at least one evaluation event in the measurement signal, - Providing (S6) at least one evaluation rule for evaluating the measurement signal, a determination event being determined on the basis of the evaluation rule when an evaluation event corresponding to the event frequency occurs - applying (S7) the at least one evaluation rule to the measurement signal, Determining (S8) a number of ascertainment events on the basis of the applied evaluation rule, - Determining (S9) the number of movements based on the number n and the determined number of detection events. Procedure according to Claim 1 , where n = 2 or n = 4 is selected for the number n. Method according to one of the Claims 1 - 2nd , whereby the amount of the at least two predetermined movement types includes at least two of the following movement states: "walking", "walking", "Nordic walking", "running", "jogging", "hopping", and / or "sprinting". Method according to one of the Claims 1 - 3rd , wherein the measurement signal is analyzed to determine a measurement signal property, in particular where a frequency analysis of the measurement signal is carried out, and the current type of movement is determined from the plurality of movement types provided on the basis of the determined measurement signal property. Method according to one of the Claims 1 - 4th , the event frequency being provided based on the current movement type determined. Method according to one of the Claims 1 - 5 , the at least one evaluation rule taking into account a deviation interval of the event frequency. Method according to one of the Claims 1 - 6 , wherein the evaluation rule is provided such that peaks of a sum signal, in particular an amount of an acceleration vector in at least two spatial directions, are evaluated in order to determine a determination event. Method according to one of the Claims 1 - 7 , wherein an inertial sensor is provided as an acceleration sensor. Procedure according to Claim 3 , wherein to provide the at least two predetermined types of movement, statistical evaluation of provided acceleration values takes place. Procedure according to Claim 9 , wherein the statistical evaluation takes place on the basis of a calculation of mean values (7) and variances (8) of the acceleration values provided, in particular wherein the acceleration values provided are smoothed exponentially. System (1) for determining periodic movements of a person, comprising at least one inertial sensor for measuring movements of a person and for providing at least one measurement signal for the measured movements, a determination device which is connected to the inertial sensor and which is designed to carry out the following steps: - Providing a plurality of different types of movement of a person, each type of movement at a different periodic Sequence of movements corresponds, and each movement type is assigned a positive integer n, at least one of the assigned numbers being greater than 1, - determining a current movement type of the person based on the measurement signal, - providing at least one evaluation event for the determined current movement type , wherein the at least one evaluation event corresponds to the nth occurrence of a movement in the periodic sequence of movements of a person, - providing an event frequency for the occurrence of the at least one evaluation event in the measurement signal, - providing at least one evaluation rule for evaluating the measurement signal, wherein a determination event is determined on the basis of the evaluation rule when an evaluation event corresponding to the event frequency occurs - application of the at least one evaluation rule to the measurement signal, - determination of a number of determination events based on the evaluation rule applied, - determination eln the number of movements based on the number n and the determined number of detection events. System according to Claim 11 , wherein the at least one inertial sensor is designed as an acceleration and / or rotation rate sensor, in particular in the form of a multi-axis sensor.

Images