[go: up one dir, main page]

EP1714112A1 - Method of sensing the motion of a solid, using an absolute measurement that is associated with a measurement calculated by double integration - Google Patents

Method of sensing the motion of a solid, using an absolute measurement that is associated with a measurement calculated by double integration

Info

Publication number
EP1714112A1
EP1714112A1 EP04816578A EP04816578A EP1714112A1 EP 1714112 A1 EP1714112 A1 EP 1714112A1 EP 04816578 A EP04816578 A EP 04816578A EP 04816578 A EP04816578 A EP 04816578A EP 1714112 A1 EP1714112 A1 EP 1714112A1
Authority
EP
European Patent Office
Prior art keywords
measurement
solid
translation
movement
rotation
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
EP04816578A
Other languages
German (de)
French (fr)
Inventor
Dominique David
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Commissariat a lEnergie Atomique et aux Energies Alternatives CEA
Original Assignee
Commissariat a lEnergie Atomique CEA
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Commissariat a lEnergie Atomique CEA filed Critical Commissariat a lEnergie Atomique CEA
Publication of EP1714112A1 publication Critical patent/EP1714112A1/en
Ceased legal-status Critical Current

Links

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Measuring devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor or mobility of a limb
    • A61B5/1113—Local tracking of patients, e.g. in a hospital or private home
    • A61B5/1114—Tracking parts of the body
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Measuring devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor or mobility of a limb
    • A61B5/1121—Determining geometric values, e.g. centre of rotation or angular range of movement
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Measuring devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor or mobility of a limb
    • A61B5/1126—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor or mobility of a limb using a particular sensing technique
    • A61B5/1127—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor or mobility of a limb using a particular sensing technique using markers
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B2562/00—Details of sensors; Constructional details of sensor housings or probes; Accessories for sensors
    • A61B2562/02—Details of sensors specially adapted for in-vivo measurements
    • A61B2562/0219—Inertial sensors, e.g. accelerometers, gyroscopes, tilt switches
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Measuring devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/107—Measuring physical dimensions, e.g. size of the entire body or parts thereof
    • A61B5/1071—Measuring physical dimensions, e.g. size of the entire body or parts thereof measuring angles, e.g. using goniometers

Definitions

  • the present invention relates to a method 5 making it possible to measure - it is also said to capture (in English "sensible") - the movement of an object or, more precisely, of a solid, that is to say to measure the displacements of this solid. It will be recalled that any displacement of a solid decomposes into a translation and a rotation (but can be limited to a simple translation or to a simple rotation).
  • the invention applies in particular to the capture of the movements of the human body. 15 It thus finds applications for example in the sports field, the medical field, the cinema, multimedia and augmented reality.
  • the invention allows movement to be captured reliably and inexpensively, even in the case of rapid movement of a person.
  • the invention completes the technique which is described in document [1] and which uses a device called an “attitude center”, comprising at least one angular position sensor (preferably at least one accelerometer and at least one magnetometer).
  • the invention makes it possible to increase the performance of this technique, in particular in the case of rapid movements.
  • various more or less efficient techniques are known for determining the displacement of a mobile object.
  • the double integration method is known in particular from acceleration measurements carried out by means of one or more accelerometers. This double integration method is implemented in positioning systems called “inertial systems" and gives good results, even in the case of rapid movements or, more precisely, movements whose speed varies rapidly.
  • the double integration of signals supplied by accelerometers is a source of positioning drift.
  • the aim of the present invention is to remedy the drawbacks of the known methods for measuring the movement of a solid, which have been mentioned above, namely the absolute measurement method and the double integration method, in order to obtain a method which is not a source of drift in positioning and can be implemented to study movements whose speed varies rapidly.
  • the dual integration method and the absolute measurement method are combined in order to readjust the measurements provided by the dual integration method, by means of the measurements provided by the absolute measurement method, these latter measurements being taken into account when the solid whose movement is measured slows down or, more precisely, when the speed of this solid varies slowly.
  • the present invention relates to a method of measuring the movement of a solid, method in which one measures at least a first translation (first degree of freedom) of this solid, this method comprising a series of stages of measurement of the acceleration of the solid and double integration of the measurements thus carried out to obtain successive values of the first translation, this method being characterized in that it further comprises a series of steps of absolute measurement of at least a second degree of freedom of the solid, this second degree of freedom being a rotation, using at least one rotation sensor, and in that this measurement of rotation is transformed into a measurement of translation and this measurement of translation is used to readjust the first translation.
  • the second degree of freedom measurement obtained at this stage can be used as an initial condition for obtaining, by double integration, the value of the first translation which follows the values previously obtained from this first translation.
  • the steps of absolute measurement and the steps of measuring the acceleration of the solid can be simultaneous, each step of absolute measurement thus taking place at the same time as a step of measuring the acceleration of the solid.
  • the transformation of the measurement of rotation into a measurement of translation uses kinetic models of the solid and / or of its movement, making it possible to establish relations between rotation and translation.
  • the rotation sensor is preferably chosen from accelerometers and magnetometers (and the second degree of freedom is therefore measured at using at least one accelerometer and / or at least one accelerometer).
  • the first translation is measured using a translation sensor which is also the rotation sensor.
  • a translation sensor which is also the rotation sensor.
  • the criterion of slowness of the movement can be the situation of a function of the standard of the acceleration of the solid below a predefined threshold.
  • the inertial sensor consists of one or more accelerometers, the accuracy of which is as high as possible. It is optionally possible to use three accelerometers each having a sensitivity axis, the respective sensitivity axes being orthogonal two by two, or an accelerometer having three axes of orthogonal sensitivity two by two.
  • the measurement process is then as follows.
  • the solid translation data is calculated by a double integration of the signals supplied by the accelerometer (s) and the rotation data is calculated from the attitude unit.
  • This attitude controller is capable of determining whether the movement in progress is fast or slow and therefore whether the values it supplies are exact or biased, for example by an evaluation of the absolute value of the amplitude of the measurements supplied by the or the accelerometers in the attitude station.
  • the data from this central attitude is used to readjust the movement of the solid.
  • the output signals from the accelerometer (s) (which are preferably of high precision) are integrated twice and thus provide a more exact response than that provided by the attitude station. It should be noted that this process does not cover the capture of movements in all its generality. However, it ideally covers the case of capturing the movements of the human body and, more generally, of the body of a vertebrate, or even of an assembly of mechanically articulated rigid segments.
  • this method is also applicable when an a priori model of the movement of the solid is known, for example in the case of a ballistic movement.
  • a priori model of the movement of the solid is known, for example in the case of a ballistic movement.
  • a method according to the invention then consists in using the double integration in the phases where the movements of the body are rapid (these phases being generally short, because the body can only carry out periodic movements) and in switching to registration mode. whenever the acceleration of body movements becomes weak.
  • Figure 1 is a schematic view of a device for implementing a method according to the invention.
  • This device makes it possible to measure the movement of a solid 2 and comprises: - one or more accelerometers 4, - one or more accelerometers 6 and / or one or more magnetometers 8, and - electronic means 10 provided for storing and processing, in accordance in the invention, the information or signals supplied by the accelerometer (s) 4 and by the accelerometer (s) 6 and / or the magnetometer (s) 8, and for storing the results of the processing.
  • the accelerometer (s) 4 as well as the accelerometer (s) 6 and / or the magnetometer (s) 8 are fixed to the solid 2 whose movement is to be measured.
  • the electronic means 10 may or may not be integral with this solid 2.
  • the electronic means 10 are therefore provided for implementing the invention by cooperating with the accelerometer (s) 4 and with the accelerometer (s) 6 and / or the magnetometer (s) 8. In particular, they cooperate with the magnetometer (s) 4 to implement a method of measurement by double integration and with the accelerometer (s) 6 and / or the magnetometer (s) 8 to implement an absolute measurement method.
  • the reference 12 symbolizes an output from the electronic means 10, in which the results of the processing can be retrieved, for example with a view to displaying these results. Examples of the invention are considered in the following, in which angular data and accelerometric data are used jointly.
  • the first example relates to a double permanent integration failed (in English "updated").
  • Each measurement point is equipped with a set of sensors, comprising from 1 to 3 accelerometers and possibly from 1 to 3 magnetometers.
  • three accelerometers are each used, each having an axis of sensitivity, they are advantageously arranged so that their respective axes of sensitivity form a triangular trihedron. It is the same when using three magnetometers each having an axis of sensitivity.
  • the data acceleration are integrated twice permanently.
  • the output signals obtained therefore result from this double integration.
  • the latter is subject to drift, the amplitude of which depends on the quality of the accelerometer or accelerometers used.
  • the data from the angular unit is acquired continuously in parallel.
  • a quality index of these angular data is also calculated. It is a function of the standard
  • This quality index is used as a criterion of slowness of the movement or, more precisely, of slowness of the variation in the speed of this movement. When the movement is sufficiently slow, which is determined for example by comparing this index to a predefined threshold and determining if the index is less than this threshold, then the angular data are used to calculate a position resulting from the movement of the solid studied.
  • the second example which is described with reference to Figure 2, relates to a method known as the "variable lever arm".
  • two rigid parts 14 and 16 are seen which are articulated by any appropriate means 18.
  • the part 14, the part 16 and their articulation 18 are respectively an arm and the corresponding forearm and elbow .
  • the forearm 16 is equipped with two assemblies 20 and 22 spaced from one another.
  • Each set has one to three accelerometers and constitutes a measurement point. When it has three accelerometers, the latter are mounted so that their axes of sensitivity form a three-dimensional trihedron. At least one of the two measurement points 20 and 22 also includes three magnetometers, the axes of sensitivity of which form an advantageously trirectangle trihedron.
  • the two sets 20 and 22 record different accelerations since the acceleration recorded by a set depends the distance from the latter to the center of rotation, namely the bend 18 in the example. The difference in accelerations is used to assess the acceleration component after eliminating the contribution of gravity.
  • the estimate of this measurement is then used in the calculation of the angles of rotation (see document [1]) by being subtracted from the total acceleration, measured by one of the two sets, which is then provided with 6 sensors (three accelerometers and three magnetometers). This gives access to a measurement of the angles, which is freed from the disturbance due to rapid movement. All these calculations are done in electronic processing means 24 which receive the signals supplied by the sensors 20 and 22.
  • the third example relates to the exploitation of a model of the movement. In this third example, we rely on the fact that certain movements are very limited as regards the number of degrees of freedom. For example, in the case of the human body, a thigh is almost limited to a single degree of freedom of rotation in the walking and running phases.
  • the movement considered can be described by a single parameter, or even by a single value of this parameter.
  • the maximum value of the measured acceleration thus makes it possible to know the entire rotation.
  • the preceding considerations are based on physiological studies which establish such results.
  • the process according to the invention, which is implemented in this third example, is then as follows. Knowing a starting position easily identifiable, because it is either a stopped position, or a reversal of rotation or translation, the maximum value of the standard of the acceleration vector is measured in the following phase, by means of three sensors whose axes of sensitivity form a trirectangle trihedron, until we identify and we know a new characteristic step. We use all of these data to extract the parameter necessary to qualify the entire movement. This step is no longer carried out in real time, since it requires knowing the whole movement, but with a (slight) delay.
  • the algorithm is as follows: (a) we take as the movement state the state calculated at the previous step (position, speed, acceleration), (b) we deduce the measurement values expected at the output of the sensors, (c) using a conventional mathematical optimization technique (for example the gradient descent method or more recent analogous methods), the initial state values of the movement are corrected, and (d) we return to step (a) until the estimated values at the output of the sensors are sufficiently close to the actual measured values.
  • a we take as the movement state the state calculated at the previous step (position, speed, acceleration), (b) we deduce the measurement values expected at the output of the sensors, (c) using a conventional mathematical optimization technique (for example the gradient descent method or more recent analogous methods), the initial state values of the movement are corrected, and (d) we return to step (a) until the estimated values at the output of the sensors are sufficiently close to the actual measured values.
  • a conventional mathematical optimization technique for example the gradient descent method or more recent analogous methods
  • FIG. 3 This example is schematically illustrated in FIG. 3. It is limited to the arm in the following but generalizable to the whole of the body.
  • the references 26, 28, 30, 32 and 34 respectively represent the shoulder, the arm, the elbow, the forearm and the hand.
  • the initial and final positions of the hand have the references 36 and 38 respectively.
  • the movement of the hand which performs a vertical translation of amplitude D, results in a rotation of angle around the elbow and possibly, depending on the amplitude of the translation, by another rotation around the shoulder. Instead of measuring D directly, we can therefore measure ⁇ . Knowing the length r of the forearm, we deduce the amplitude D of the translation.
  • This technique according to the invention has the advantage of only being based on measurements absolute: it is therefore free from drifts.
  • this technique uses, among other things, one or more accelerometers (attached to the forearm but not shown) to measure the angle ⁇ .
  • accelerometers Attached to the forearm but not shown
  • the accelerometer (s) additionally measure the acceleration resulting from such a movement, so that the angle measurement is distorted.
  • a technique known from document [1] partially solves this problem. It consists in reducing the contribution of the accelerometer (s) in favor of the magnetometer (s) in the calculation of the angle (s). But this technique is only partially effective and depends on the movement performed.
  • a technique proposed in the present invention is capable of supplementing the previous one and is not limited by the type of movement allowed.
  • the displacement of the sensor is calculated by double integration, taking as a starting point the state of the mobile solid (the arm in the example) at the start of the rapid phase, we simultaneously correct any possible drift, by merging the agnometric data, and as soon as the movement slows down, we return to absolute mode.
  • the fusion of magnetometric data consists in using, in conjunction with double integration, an estimate of rapid movement using only magnetometers. The latter technique has been described above (reducing the contribution of the accelerometer (s) in favor of the magnetometer (s) in the calculation of the angle (s)).
  • double integration provides in theory the complete movement, but it is subject to drift.
  • magnetometers provide a partial estimate (excluding rotations around the axis of the earth's field) but not subject to drift.
  • a possible fusion consists in estimating the movement from the double integration, in deducing therefrom estimated magnetic measurements, and in using the difference between the latter and the real magnetic measurements to correct the motion estimated by a technique of “gradient descent” type. ".
  • a fusion algorithm can be implemented between the double integration method and the absolute measurement method, by passing from one to the other not discontinuously but gradually, gradually decreasing the absolute contribution of accelerometers and in gradually increasing the influence of double integration when the movement accelerates, and vice versa during the deceleration phase.
  • the state of the mobile is estimated (position, speed, acceleration) from the last known state and the double integration of high precision accelerometers.
  • the estimated movement is corrected by applying a "gradient descent" type method.
  • the correction in question is parameterized: it is all the more important the slower the movement, the criterion being for example the ratio of the norm of the vector acceleration to the norm of g, ratio which was mentioned above.
  • the criterion being for example the ratio of the norm of the vector acceleration to the norm of g, ratio which was mentioned above.
  • the invention has all the advantages of the technique which is described in document [1]: it can be implemented inexpensively, it does not require any external equipment, such as magnetic sources or cameras, and it can be implemented with robust algorithms. In addition, the invention leads to reliable measurements even in the case of rapid movements.
  • the present invention can be implemented with an attitude unit whose angular precision is less than or equal to 1 ° and with accelerometers at least 10 bits (advantageously from 14 to 16 bits). It is specified that the criterion of "slowness of the movement", which we mentioned above, is a function of the precision that we want to obtain on the movement.
  • An object of the invention being to separate the acceleration of the solid from the acceleration of gravity, as long as the standard of the acceleration of the solid remains below - Il g II (little different from lm / s 2 ), the movement will be considered slow and the process will lead to acceptable accuracy.

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Molecular Biology (AREA)
  • Dentistry (AREA)
  • Biophysics (AREA)
  • Pathology (AREA)
  • Engineering & Computer Science (AREA)
  • Biomedical Technology (AREA)
  • Physiology (AREA)
  • Medical Informatics (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Geometry (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
  • Length Measuring Devices With Unspecified Measuring Means (AREA)
  • Navigation (AREA)

Abstract

The invention relates to a method of sensing the motion of a solid, using an absolute measurement that is associated with a measurement calculated by double integration, which is intended, for example, for human body motion sensing. The inventive method comprises: a series of steps involving the measurement of the acceleration of the solid (2) and the double integration of said measurements in order to obtain successive values of a first translation of the solid, and a series of steps involving the absolute measurement of at least one second degree of freedom of the solid, namely a rotation. According to the invention, the rotation measurement is transformed into a translation measurement and said translation measurement is used to update the first translation.

Description

PROCEDE DE CAPTURE DU MOUVEMENT D'UN SOLIDE, UTILISANT UNE MESURE ABSOLUE ASSOCIEE A UNE MESURE PAR DOUBLE INTEGRATIONP R OCEDE OF CAPTURE OF THE MOVEMENT OF A SOLID, USING AN ABSOLUTE MEASUREMENT ASSOCIATED WITH A MEASUREMENT BY DOUBLE INTEGRATION

DESCRIPTIONDESCRIPTION

DOMAINE TECHNIQUE La présente invention concerne un procédé 5 permettant de mesurer - on dit aussi capturer (en anglais « sensé ») - le mouvement d'un objet ou, plus précisément, d'un solide, c'est-à-dire de mesurer les déplacements de ce solide. On rappelle que tout déplacement d'un 10 solide se décompose en une translation et une rotation (mais peut se limiter à une simple translation ou à une simple rotation) . L'invention s'applique en particulier à la capture des mouvements du corps humain. 15 Elle trouve ainsi des applications par exemple dans le domaine sportif, le domaine médical, le cinéma, le multimédia et la réalité augmentée. L'invention permet de capturer un mouvement de façon fiable et peu coûteuse, même dans le cas d'un 20 mouvement rapide d'une personne.TECHNICAL FIELD The present invention relates to a method 5 making it possible to measure - it is also said to capture (in English "sensible") - the movement of an object or, more precisely, of a solid, that is to say to measure the displacements of this solid. It will be recalled that any displacement of a solid decomposes into a translation and a rotation (but can be limited to a simple translation or to a simple rotation). The invention applies in particular to the capture of the movements of the human body. 15 It thus finds applications for example in the sports field, the medical field, the cinema, multimedia and augmented reality. The invention allows movement to be captured reliably and inexpensively, even in the case of rapid movement of a person.

ETAT DE LA TECHNIQUE ANTERIEURE On se reportera au document suivant :STATE OF THE PRIOR ART Reference is made to the following document:

25 [1] WO 03/085357A, demande internationale numéro PCT/FR03/01025, déposée le 2 avril 2003, « Dispositif de capture des mouvements de rotation d'un solide », invention de Dominique David et Yanis Caritu.25 [1] WO 03 / 085357A, international application number PCT / FR03 / 01025, filed on April 2, 2003, "Device for capturing the rotational movements of a solid", invention of Dominique David and Yanis Caritu.

L'invention complète la technique qui est décrite dans le document [1] et qui utilise un dispositif appelé « centrale d'attitude », comprenant au moins un capteur de position angulaire (de préférence au moins un accéléromètre et au moins un magnétomètre) . L'invention permet d'accroître les performances de cette technique, en particulier dans le cas de mouvements rapides . On connaît en effet diverses techniques plus ou moins performantes pour déterminer le déplacement d'un objet mobile. On connaît en particulier la méthode de double intégration à partir de mesures d'accélération, effectuées au moyen d'un ou de plusieurs accéléromètres . Cette méthode de double intégration est mise en œuvre dans les systèmes de positionnement appelés « systèmes inertiels » et donne de bons résultats, même dans le cas de mouvements rapides ou, plus précisément, de mouvements dont la vitesse varie rapidement. Cependant, la double intégration de signaux fournis par des accéléromètres est une source de dérive du positionnement. Afin de limiter cette dérive, en particulier dans le domaine de l'aviation ou le domaine spatial, on est amené à utiliser des accéléromètres qui sont très performants mais malheureusement très coûteux. On connaît en outre la méthode de mesure absolue d'un mouvement, à partir d'un ou de plusieurs accéléromètres et d'un ou de plusieurs magnétomètres . Cette méthode n'est pas source de dérive mais permet seulement de mesurer le mouvement d'un objet dont la vitesse varie lentement.The invention completes the technique which is described in document [1] and which uses a device called an “attitude center”, comprising at least one angular position sensor (preferably at least one accelerometer and at least one magnetometer). The invention makes it possible to increase the performance of this technique, in particular in the case of rapid movements. In fact, various more or less efficient techniques are known for determining the displacement of a mobile object. The double integration method is known in particular from acceleration measurements carried out by means of one or more accelerometers. This double integration method is implemented in positioning systems called "inertial systems" and gives good results, even in the case of rapid movements or, more precisely, movements whose speed varies rapidly. However, the double integration of signals supplied by accelerometers is a source of positioning drift. In order to limit this drift, in particular in the field of aviation or the space field, it is necessary to use accelerometers which are very efficient but unfortunately very expensive. The method of absolute measurement of a movement is also known, from one or more accelerometers and from one or more magnetometers. This method is not a source of drift but only makes it possible to measure the movement of an object whose speed varies slowly.

EXPOSE DE L'INVENTION La présente invention a pour but de remédier aux inconvénients des méthodes connues de mesure du mouvement d'un solide, que l'on a mentionnées plus haut, à savoir la méthode de mesure absolue et la méthode de double intégration, en vue d'obtenir une méthode qui n'est pas source de dérive du positionnement et peut être mise en œuvre pour étudier des mouvements dont la vitesse varie rapidement. Selon un aspect particulier de l'invention, on combine la méthode de double intégration et la méthode de mesure absolue afin de recaler les mesures fournies par la méthode de double intégration, au moyen des mesures fournies par la méthode de mesure absolue, ces dernières mesures étant prises en compte lorsque le solide dont on mesure le mouvement ralentit ou, plus précisément, lorsque la vitesse de ce solide varie lentement. De façon précise, la présente invention a pour objet un procédé de mesure du mouvement d'un solide, procédé dans lequel on mesure au moins une première translation (premier degré de liberté) de ce solide, ce procédé comprenant une série d'étapes de mesure de l'accélération du solide et de double intégration des mesures ainsi effectuées pour obtenir des valeurs successives de la première translation, ce procédé étant caractérisé en ce qu'il comprend en outre une série d'étapes de mesure absolue d'au moins un deuxième degré de liberté du solide, ce deuxième degré de liberté étant une rotation, à l'aide d'au moins un capteur de rotation, et en ce que l'on transforme cette mesure de rotation en une mesure de translation et cette mesure de translation est utilisée pour recaler la première translation. Dans la présente invention, on peut utiliser la mesure du deuxième degré de liberté, obtenue à cette étape, en tant que condition initiale pour obtenir, par double intégration, la valeur de la première translation qui suit les valeurs précédemment obtenues de cette première translation. Dans la présente invention, les étapes de mesure absolue et les étapes de mesure de l'accélération du solide peuvent être simultanées, chaque étape de mesure absolue ayant ainsi lieu en même temps qu'une étape de mesure de l'accélération du solide. De préférence, la transformation de la mesure de rotation en une mesure de translation utilise des modèles cinétiques du solide et/ou de son mouvement, permettant d'établir des relations entre la rotation et la translation. Le capteur de rotation est de préférence choisi parmi les accéléromètres et les magnétomètres (et l'on mesure donc le deuxième degré de liberté à l'aide d'au moins un accélëro ètre et/ou d'au moins un agnétomètre) . Selon un mode de réalisation particulier de l'invention, la première translation est mesurée à l'aide d'un capteur de translation qui est aussi le capteur de rotation. De préférence, on choisit un critère de lenteur du mouvement (plus précisément un critère de variation lente de la vitesse du solide) et, si le mouvement satisfait à ce critère après l'une des étapes de mesure du deuxième degré de liberté, on utilise la mesure du deuxième degré de liberté, obtenue à cette étape, pour recaler (en anglais « update ») la première translation. Le critère de lenteur du mouvement peut être la situation d'une fonction de la norme de l'accélération du solide en deçà d'un seuil prédéfiniSUMMARY OF THE INVENTION The aim of the present invention is to remedy the drawbacks of the known methods for measuring the movement of a solid, which have been mentioned above, namely the absolute measurement method and the double integration method, in order to obtain a method which is not a source of drift in positioning and can be implemented to study movements whose speed varies rapidly. According to a particular aspect of the invention, the dual integration method and the absolute measurement method are combined in order to readjust the measurements provided by the dual integration method, by means of the measurements provided by the absolute measurement method, these latter measurements being taken into account when the solid whose movement is measured slows down or, more precisely, when the speed of this solid varies slowly. Specifically, the present invention relates to a method of measuring the movement of a solid, method in which one measures at least a first translation (first degree of freedom) of this solid, this method comprising a series of stages of measurement of the acceleration of the solid and double integration of the measurements thus carried out to obtain successive values of the first translation, this method being characterized in that it further comprises a series of steps of absolute measurement of at least a second degree of freedom of the solid, this second degree of freedom being a rotation, using at least one rotation sensor, and in that this measurement of rotation is transformed into a measurement of translation and this measurement of translation is used to readjust the first translation. In the present invention, the second degree of freedom measurement obtained at this stage can be used as an initial condition for obtaining, by double integration, the value of the first translation which follows the values previously obtained from this first translation. In the present invention, the steps of absolute measurement and the steps of measuring the acceleration of the solid can be simultaneous, each step of absolute measurement thus taking place at the same time as a step of measuring the acceleration of the solid. Preferably, the transformation of the measurement of rotation into a measurement of translation uses kinetic models of the solid and / or of its movement, making it possible to establish relations between rotation and translation. The rotation sensor is preferably chosen from accelerometers and magnetometers (and the second degree of freedom is therefore measured at using at least one accelerometer and / or at least one accelerometer). According to a particular embodiment of the invention, the first translation is measured using a translation sensor which is also the rotation sensor. Preferably, one chooses a criterion of slowness of the movement (more precisely a criterion of slow variation of the speed of the solid) and, if the movement satisfies this criterion after one of the steps of measuring the second degree of freedom, one uses the measurement of the second degree of freedom, obtained at this stage, to readjust (in English "update") the first translation. The criterion of slowness of the movement can be the situation of a function of the standard of the acceleration of the solid below a predefined threshold.

(en anglais « predetermined ») . Cette fonction peut être cette norme elle- même .(in English "predetermined"). This function can be this standard itself.

BREVE DESCRIPTION DES DESSINS La présente invention sera mieux comprise à la lecture de la description d'exemples de réalisation donnés ci-après, à titre purement indicatif et nullement limitatif, en faisant référence aux dessins annexés, sur lesquels : - la figure 1 illustre schématiquement un dispositif pour la mise en œuvre d'un exemple du procédé objet de l'invention, et les figures 2 et 3 illustrent schématiquement des exemples de l'invention.BRIEF DESCRIPTION OF THE DRAWINGS The present invention will be better understood on reading the description of exemplary embodiments given below, by way of purely indicative and in no way limiting, with reference to the appended drawings, in which: - Figure 1 schematically illustrates a device for implementing an example of the method which is the subject of the invention, and Figures 2 and 3 schematically illustrate examples of the invention.

EXPOSÉ DÉTAILLÉ DE MODES DE RÉALISATION PARTICULIERS Dans un exemple de l'invention, on cherche à déterminer en permanence la position d'un objet, plus précisément d'un solide, en mesurant les valeurs des six degrés de liberté de ce solide. Le dispositif (centrale d'attitude) décrit dans le document [1] permet de mesurer les trois degrés de liberté angulaires du solide. La présente invention permet de compléter ce dispositif connu par un dispositif inertiel, qui en étend donc les capacités. Une originalité de la présente invention réside dans le type de couplage qu'elle propose, à savoir une technique utilisant un capteur absolu et une technique utilisant un capteur inertiel. Le capteur absolu est la centrale d'attitude qui fournit des mesures absolues de positionnement angulaire du solide. Ces mesures sont exactes lorsque le solide est au repos ; mais elles sont susceptibles d'être entachées d'erreurs lorsque le solide subit des accélérations, les erreurs étant d'autant plus importantes que les accélérations sont plus fortes. Le capteur inertiel est constitué par un ou plusieurs accéléromètres dont la précision est aussi grande que possible. On peut éventuellement utiliser trois accéléromètres ayant chacun un axe de sensibilité, les axes de sensibilité respectifs étant orthogonaux deux à deux, ou un accéléromètre ayant trois axes de sensibilité orthogonaux deux à deux. Le procédé de mesure est alors le suivant. On calcule les données de translation du solide par une double intégration des signaux fournis par le ou les accéléromètres et l'on calcule les données de rotation à partir de la centrale d'attitude. Cette centrale d'attitude est capable de déterminer si le mouvement en cours est rapide ou lent et donc si les valeurs qu'elle fournit sont exactes ou biaisees, par exemple par une évaluation de la valeur absolue de l'amplitude des mesures fournies par le ou les accéléromètres que comporte la centrale d'attitude. Dans les phases de mouvement lent, les données de cette centrale d'attitude sont utilisées pour recaler le mouvement du solide. Dans les phases de mouvement rapide, les signaux de sortie du ou des accéléromètres (qui sont de préférence de haute précision) sont intégrés deux fois et fournissent ainsi une réponse plus exacte que celle qui est fournie par la centrale d'attitude. Il convient de noter que ce procédé ne couvre pas la capture de mouvements dans toute sa généralité. Il couvre par contre idéalement le cas de la capture des mouvements du corps humain et, plus généralement, du corps d'un vertébré, voire d'un assemblage de segments rigides mécaniquement articulés . Dans certains cas, ce procédé est également applicable quand un modèle a priori du mouvement du solide est connu, par exemple dans le cas d'un mouvement balistique. Dans le cas du corps humain, on connaît la posture complète du corps grâce à une conjonction de centrales d'attitude disposées sur les segments osseux. Cette mesure est entachée d'erreur en cas de mouvements rapides du corps . Un procédé conforme à l'invention consiste alors à utiliser la double intégration dans les phases où les mouvements du corps sont rapides (ces phases étant en général courtes, car le corps ne peut effectuer que des mouvements périodiques) et à passer en mode de recalage chaque fois que les accélérations des mouvements du corps deviennent faibles . La figure 1 est une vue schématique d'un dispositif pour la mise en œuvre d'un procédé conforme à l'invention. Ce dispositif permet de mesurer le mouvement d'un solide 2 et comprend : - un ou plusieurs accéléromètres 4, - un ou plusieurs accéléromètres 6 et/ou un ou plusieurs magnétomètres 8, et - des moyens électroniques 10 prévus pour mémoriser et traiter, conformément à l'invention, les informations ou signaux fournis par le ou les accéléromètres 4 et par le ou les accéléromètres 6 et/ou le ou les magnétomètres 8, et pour mémoriser les résultats du traitement. Le ou les accéléromètres 4 ainsi que le ou les accéléromètres 6 et/ou le ou les magnétomètres 8 sont fixés au solide 2 dont on veut mesurer le mouvement. Les moyens électroniques 10 peuvent être solidaires, ou non, de ce solide 2. Les moyens électroniques 10 sont donc prévus pour mettre en œuvre l'invention en coopérant avec le ou les accéléromètres 4 et avec le ou les accéléromètres 6 et/ou le ou les magnétomètres 8. En particulier, ils coopèrent avec le ou les magnétomètres 4 pour mettre en œuvre une méthode de mesure par double intégration et avec le ou les accéléromètres 6 et/ou le ou les magnétomètres 8 pour mettre en œuvre une méthode de mesure absolue . Sur la figure 1, la référence 12 symbolise une sortie des moyens électroniques 10, en laquelle on peut récupérer les résultats du traitement, par exemple en vue d'un affichage (en anglais « display ») de ces résultats . On considère dans ce qui suit des exemples de l'invention, dans lesquels on exploite conjointement des données angulaires et des données accélérométriques . Le premier exemple se rapporte à une double intégration permanente recalée (en anglais « updated ») . Chaque point de mesure est équipé d'un ensemble de capteurs, comprenant de 1 à 3 accéléromètres et éventuellement de 1 à 3 magnétomètres . Lorsque l'on utilise trois accéléromètres ayant chacun un axe de sensibilité, on les dispose avantageusement de façon que leurs axes de sensibilité respectifs forment un trièdre trirectangle . Il en est de même lorsque l'on utilise trois magnétomètres ayant chacun un axe de sensibilité. Dans ce premier exemple, les données d'accélération sont intégrées deux fois en permanence.DETAILED PRESENTATION OF PARTICULAR EMBODIMENTS In an example of the invention, it is sought to permanently determine the position of an object, more precisely of a solid, by measuring the values of the six degrees of freedom of this solid. The device (central attitude) described in document [1] makes it possible to measure the three angular degrees of freedom of the solid. The present invention makes it possible to supplement this known device with an inertial device, which therefore extends its capacities. An originality of the present invention lies in the type of coupling which it offers, namely a technique using an absolute sensor and a technique using an inertial sensor. The absolute sensor is the attitude controller which provides absolute measurements of the angular positioning of the solid. These measurements are accurate when the solid is at rest; but they are likely to be marred by errors when the solid undergoes accelerations, the errors being all the more important as the accelerations are stronger. The inertial sensor consists of one or more accelerometers, the accuracy of which is as high as possible. It is optionally possible to use three accelerometers each having a sensitivity axis, the respective sensitivity axes being orthogonal two by two, or an accelerometer having three axes of orthogonal sensitivity two by two. The measurement process is then as follows. The solid translation data is calculated by a double integration of the signals supplied by the accelerometer (s) and the rotation data is calculated from the attitude unit. This attitude controller is capable of determining whether the movement in progress is fast or slow and therefore whether the values it supplies are exact or biased, for example by an evaluation of the absolute value of the amplitude of the measurements supplied by the or the accelerometers in the attitude station. In the slow movement phases, the data from this central attitude is used to readjust the movement of the solid. In the rapid movement phases, the output signals from the accelerometer (s) (which are preferably of high precision) are integrated twice and thus provide a more exact response than that provided by the attitude station. It should be noted that this process does not cover the capture of movements in all its generality. However, it ideally covers the case of capturing the movements of the human body and, more generally, of the body of a vertebrate, or even of an assembly of mechanically articulated rigid segments. In certain cases, this method is also applicable when an a priori model of the movement of the solid is known, for example in the case of a ballistic movement. In the case of the human body, we know the full posture of the body thanks to a conjunction of attitude centers located on the bone segments. This measurement is vitiated by error in the event of rapid movements of the body. A method according to the invention then consists in using the double integration in the phases where the movements of the body are rapid (these phases being generally short, because the body can only carry out periodic movements) and in switching to registration mode. whenever the acceleration of body movements becomes weak. Figure 1 is a schematic view of a device for implementing a method according to the invention. This device makes it possible to measure the movement of a solid 2 and comprises: - one or more accelerometers 4, - one or more accelerometers 6 and / or one or more magnetometers 8, and - electronic means 10 provided for storing and processing, in accordance in the invention, the information or signals supplied by the accelerometer (s) 4 and by the accelerometer (s) 6 and / or the magnetometer (s) 8, and for storing the results of the processing. The accelerometer (s) 4 as well as the accelerometer (s) 6 and / or the magnetometer (s) 8 are fixed to the solid 2 whose movement is to be measured. The electronic means 10 may or may not be integral with this solid 2. The electronic means 10 are therefore provided for implementing the invention by cooperating with the accelerometer (s) 4 and with the accelerometer (s) 6 and / or the magnetometer (s) 8. In particular, they cooperate with the magnetometer (s) 4 to implement a method of measurement by double integration and with the accelerometer (s) 6 and / or the magnetometer (s) 8 to implement an absolute measurement method. In FIG. 1, the reference 12 symbolizes an output from the electronic means 10, in which the results of the processing can be retrieved, for example with a view to displaying these results. Examples of the invention are considered in the following, in which angular data and accelerometric data are used jointly. The first example relates to a double permanent integration failed (in English "updated"). Each measurement point is equipped with a set of sensors, comprising from 1 to 3 accelerometers and possibly from 1 to 3 magnetometers. When three accelerometers are each used, each having an axis of sensitivity, they are advantageously arranged so that their respective axes of sensitivity form a triangular trihedron. It is the same when using three magnetometers each having an axis of sensitivity. In this first example, the data acceleration are integrated twice permanently.

Les signaux de sortie obtenus résultent donc de cette double intégration. Cependant, on sait que cette dernière est sujette à une dérive dont l'amplitude dépend de la qualité du ou des accéléromètres utilisés. Pour remédier à cet inconvénient, conformément à l'invention, on acquiert en parallèle, en permanence, les données de la centrale angulaireThe output signals obtained therefore result from this double integration. However, we know that the latter is subject to drift, the amplitude of which depends on the quality of the accelerometer or accelerometers used. To overcome this drawback, in accordance with the invention, the data from the angular unit is acquired continuously in parallel.

(accéléromètre (s) et éventuellement magnétomètre (s) ) . Un indice de qualité de ces données angulaires est également calculé. Il s'agit d'une fonction de la norme ||a|| du vecteur accélération par rapport à la norme de l'accélération g de la pesanteur, mesurée au repos, fonction qui peut être par exemple le rapport 11 a| | / 11 g! I . Cet indice de qualité est utilisé en tant que critère de lenteur du mouvement ou, plus précisément, de lenteur de la variation de la vitesse de ce mouvement . Lorsque le mouvement est suffisamment lent, ce que l'on détermine par exemple en comparant cet indice à un seuil prédéfini et en déterminant si l'indice est inférieur à ce seuil, alors les données angulaires sont utilisées pour calculer une position résultant du mouvement du solide étudié. Cette position sert alors de position de départ pour la période de double intégration suivante. Si le mouvement demeure quasi-statique, un recalage est effectué au bout d'un intervalle de temps défini par la précision que l'on attend des mesures, de façon que l'estimation de la dérive résultant de la double intégration demeure inférieure, dans cet intervalle de temps, à la précision voulue. Le deuxième exemple, que l'on décrit en faisant référence à la figure 2, se rapporte à une méthode dite du « bras de levier variable » . Sur la figure 2, on voit deux pièces rigides 14 et 16 qui sont articulées par tout moyen approprié 18. Par exemple, la pièce 14, la pièce 16 et leur articulation 18 sont respectivement un bras et l'avant-bras et le coude correspondants. Conformément à l'invention, l'avant-bras 16 est équipé de deux ensembles 20 et 22 espacés l'un de 1 ' autre . Chaque ensemble comporte de un à trois accéléromètres et constitue un point de mesure. Lorsqu'il comporte trois accéléromètres, ces derniers sont montés de façon que leurs axes de sensibilité forment un trièdre trirectangle . L'un au moins des deux points de mesure 20 et 22 comporte également trois magnétomètres dont les axes de sensibilité forment un trièdre avantageusement trirectangle. Lors d'un mouvement de rotation de l'avant- bras 16, du genre de celui qui est symbolisé par la flèche R sur la figure 2, les deux ensembles 20 et 22 enregistrent des accélérations différentes puisque l'accélération enregistrée par un ensemble dépend de la distance de ce dernier au centre de rotation, à savoir le coude 18 dans l'exemple. La différence des accélérations est utilisée pour évaluer la composante d'accélération après avoir éliminé la contribution de la pesanteur. L'estimation de cette mesure est ensuite utilisée dans le calcul des angles de rotation (voir document [1]) en étant soustraite de l'accélération totale, mesurée par l'un des deux ensembles, qui est alors pourvu de 6 capteurs (trois accéléromètres et trois magnétomètres) . On accède ainsi à une mesure des angles, qui est débarrassée de la perturbation due à un mouvement rapide. Tous ces calculs se font dans des moyens électroniques de traitement 24 qui reçoivent les signaux fournis par les capteurs 20 et 22. Le troisième exemple se rapporte à l'exploitation d'un modèle du mouvement. Dans ce troisième exemple, on s'appuie sur le fait que certains mouvements sont très limités en ce qui concerne le nombre de degrés de liberté. Par exemple, dans le cas du corps humain, une cuisse est quasiment limitée à un unique degré de liberté de rotation dans les phases de marche et de course. Dans ce cas, le mouvement considéré peut être décrit par un seul paramètre, voire par une seule valeur de ce paramètre. La valeur maximale de l'accélération mesurée permet ainsi de connaître l'ensemble de la rotation. Les considérations précédentes s'appuient sur des études physiologiques qui établissent de tels résultats . Le procédé conforme à l'invention, que l'on met en œuvre dans ce troisième exemple, est alors le suivant. Connaissant une position de départ facilement identifiable, parce qu'il s'agit soit d'une position arrêtée, soit d'un rebroussement de rotation ou de translation, on mesure la valeur maximale de la norme du vecteur accélération dans la phase qui suit, au moyen de trois capteurs dont les axes de sensibilité forment un trièdre trirectangle, jusqu'à ce que l'on identifie et l'on connaisse une nouvelle étape caractéristique . On exploite l'ensemble de ces données pour en extraire le paramètre nécessaire à la qualification de l'ensemble du mouvement. Cette étape n'est plus effectuée en temps réel, puisqu'elle nécessite de connaître l'ensemble du mouvement, mais avec un (léger) différé .(accelerometer (s) and possibly magnetometer (s)). A quality index of these angular data is also calculated. It is a function of the standard || a || of the vector acceleration compared to the norm of the acceleration g of gravity, measured at rest, a function which can be for example the ratio 11 a | | / 11 g! I. This quality index is used as a criterion of slowness of the movement or, more precisely, of slowness of the variation in the speed of this movement. When the movement is sufficiently slow, which is determined for example by comparing this index to a predefined threshold and determining if the index is less than this threshold, then the angular data are used to calculate a position resulting from the movement of the solid studied. This position then serves as the starting position for the next double integration period. If the movement remains quasi-static, a readjustment is carried out at the end of a time interval defined by the precision expected from the measurements, so that the estimate of the drift resulting from the double integration remains lower, in this time interval, than the desired precision. The second example, which is described with reference to Figure 2, relates to a method known as the "variable lever arm". In FIG. 2, two rigid parts 14 and 16 are seen which are articulated by any appropriate means 18. For example, the part 14, the part 16 and their articulation 18 are respectively an arm and the corresponding forearm and elbow . According to the invention, the forearm 16 is equipped with two assemblies 20 and 22 spaced from one another. Each set has one to three accelerometers and constitutes a measurement point. When it has three accelerometers, the latter are mounted so that their axes of sensitivity form a three-dimensional trihedron. At least one of the two measurement points 20 and 22 also includes three magnetometers, the axes of sensitivity of which form an advantageously trirectangle trihedron. During a rotational movement of the forearm 16, of the kind which is symbolized by the arrow R in FIG. 2, the two sets 20 and 22 record different accelerations since the acceleration recorded by a set depends the distance from the latter to the center of rotation, namely the bend 18 in the example. The difference in accelerations is used to assess the acceleration component after eliminating the contribution of gravity. The estimate of this measurement is then used in the calculation of the angles of rotation (see document [1]) by being subtracted from the total acceleration, measured by one of the two sets, which is then provided with 6 sensors (three accelerometers and three magnetometers). This gives access to a measurement of the angles, which is freed from the disturbance due to rapid movement. All these calculations are done in electronic processing means 24 which receive the signals supplied by the sensors 20 and 22. The third example relates to the exploitation of a model of the movement. In this third example, we rely on the fact that certain movements are very limited as regards the number of degrees of freedom. For example, in the case of the human body, a thigh is almost limited to a single degree of freedom of rotation in the walking and running phases. In this case, the movement considered can be described by a single parameter, or even by a single value of this parameter. The maximum value of the measured acceleration thus makes it possible to know the entire rotation. The preceding considerations are based on physiological studies which establish such results. The process according to the invention, which is implemented in this third example, is then as follows. Knowing a starting position easily identifiable, because it is either a stopped position, or a reversal of rotation or translation, the maximum value of the standard of the acceleration vector is measured in the following phase, by means of three sensors whose axes of sensitivity form a trirectangle trihedron, until we identify and we know a new characteristic step. We use all of these data to extract the parameter necessary to qualify the entire movement. This step is no longer carried out in real time, since it requires knowing the whole movement, but with a (slight) delay.

On donne dans ce qui suit d'autres exemples de 1 ' invention . On rappelle d'abord que les techniques de mesure de déplacement, qui sont fondées sur les capteurs inertiels, souffrent toutes du même défaut, à savoir d'une dérive des mesures, provenant de la double intégration de bruits d'origines diverses (notamment des bruits électroniques et des bruits physiologiques) , bruits qui s'ajoutent au signal à mesurer. Selon un aspect de la présente invention, on résout ce problème à l'aide d'une technique fondée sur un dispositif connu par le document [1] , permettant de mesurer des angles de façon absolue, au moyen de capteurs d'inclinaison et de capteurs de champ magnétique . L' originalité de cette technique repose sur la mise en oeuvre des trois modes opératoires suivants : mesure absolue d'angles au moyen d'un ou de plusieurs accéléromètres et/ou d'un ou de plusieurs magnétomètres, utilisation de modèles comportementauxIn the following, other examples of the invention are given. First of all, it is recalled that the displacement measurement techniques, which are based on inertial sensors, all suffer from the same defect, namely a drift in the measurements, originating from the double integration of noises of various origins (in particular electronic noises and physiological noises), noises which are added to the signal to be measured. According to one aspect of the present invention, this problem is solved using a technique based on a device known from document [1], making it possible to measure angles absolutely, by means of inclination and magnetic field sensors. The originality of this technique rests on the implementation of the following three operating modes: absolute measurement of angles by means of one or more accelerometers and / or one or more magnetometers, use of behavioral models

(voir l'exemple ci-dessous du corps humain), permettant de relier des angles, qui sont mesurés de façon absolue, à des translations effectives dans l'espace du solide mobile (par exemple une main) , et calcul de déplacements par double intégration de signaux issus d' accéléromètres ayant de préférence une haute précision. A la mise en oeuvre conjointe de ces trois modes opératoires, il est en outre préférable d'ajouter une méthode appropriée de fusion des données recueillies, qui est exposée ci-après. Cette méthode est en fait analogue au procédé de fusion de données décrit dans le document [1] . La complexité de ce cas provient du fait que, ici, on autorise un mouvement suffisamment rapide pour qu'il introduise une composante d'accélération se superposant à la gravité. Cette composante d'accélération ajoute 3 inconnues supplémentaires (selon les 3 axes) . Mais les accéléromètres de haute précision apportent aussi une information supplémentaire. L'algorithme est le suivant : (a) on prend comme état du mouvement l'état calculé au pas précédent (position, vitesse, accélération) , (b) on en déduit les valeurs de mesure attendues en sortie des capteurs, (c) à l'aide d'une technique d'optimisation mathématique classique (par exemple la méthode de descente de gradient ou des méthode analogues plus récentes), on corrige les valeurs d'état initiales du mouvement , et (d) on retourne à l'étape (a) jusqu'à ce que les valeurs estimées en sortie des capteurs soient suffisamment proches des valeurs réelles mesurées. On donne dans ce qui suit un exemple de l'invention, relatif à la capture des mouvements du corps humain, à six degrés de liberté.(see the example below of the human body), allowing to connect angles, which are measured in an absolute way, to effective translations in space of the mobile solid (for example a hand), and calculation of displacements by double integration of signals from accelerometers preferably having high precision. To the joint implementation of these three operating modes, it is also preferable to add an appropriate method of merging the data collected, which is set out below. This method is in fact analogous to the data fusion process described in document [1]. The complexity of this case comes from the fact that, here, a movement is allowed to be fast enough for it to introduce an acceleration component superimposed on gravity. This acceleration component adds 3 additional unknowns (along the 3 axes). But high precision accelerometers also provide additional information. The algorithm is as follows: (a) we take as the movement state the state calculated at the previous step (position, speed, acceleration), (b) we deduce the measurement values expected at the output of the sensors, (c) using a conventional mathematical optimization technique (for example the gradient descent method or more recent analogous methods), the initial state values of the movement are corrected, and (d) we return to step (a) until the estimated values at the output of the sensors are sufficiently close to the actual measured values. An example of the invention is given in the following, relating to the capture of the movements of the human body, at six degrees of freedom.

Cet exemple est schématiquement illustré par la figure 3. Il est limité au bras dans ce qui suit mais généralisable à l'ensemble du corps. Sur la figure 3, les références 26, 28, 30, 32 et 34 représentent respectivement l'épaule, le bras, le coude, l'avant-bras et la main. Les positions initiale et finale de la main ont respectivement les références 36 et 38. Le mouvement de la main, qui effectue une translation verticale d'amplitude D, se traduit par une rotation d'angle autour du coude et éventuellement, selon l'amplitude de la translation, par une autre rotation autour de l'épaule. Au lieu de mesurer directement D, on peut donc mesurer α. Connaissant la longueur r de l'avant- bras, on en déduit l'amplitude D de la translation . Cette technique conforme à l'invention présente l'avantage de ne reposer que sur des mesures absolues : elle est donc exempte de dérives. Il convient de noter que cette technique utilise entre autres un ou plusieurs accéléromètres (fixés à l'avant-bras mais non représentés) pour mesurer l'angle α. Cela suppose que de tels capteurs mesurent la pesanteur et permettant donc de connaître leurs inclinaisons respectives par rapport à la verticale. Dans le cas d'un mouvement rapide, le ou les accéléromètres mesurent en plus l'accélération résultant d'un tel mouvement, de sorte que la mesure d'angle est faussée. Une technique connue par le document [1] permet de résoudre partiellement ce problème. Elle consiste à diminuer la contribution du ou des accéléromètres au profit du ou des magnétomètres dans le calcul du ou des angles. Mais cette technique n'est que partiellement efficace et dépend du mouvement effectué. Une technique proposée dans la présente invention est capable de compléter la précédente et n'est pas limitée par le type de mouvement permis. Cette technique est exposée ci-après. Dès que l'on détecte un mouvement rapide (il suffit pour ce faire de calculer la norme du vecteur accélération) on calcule par double intégration le déplacement du capteur, en prenant pour point de départ l'état du solide mobile (l'avant-bras dans l'exemple) lors du début de la phase rapide, on corrige simultanément une éventuelle dérive, en fusionnant les données agnétométriques, et dès que le mouvement ralentit, on repasse en mode absolu. On fait ainsi disparaître une éventuelle dérive qui est susceptible d'apparaître pendant la phase d'intégration. La fusion des données magnétométriques consiste à utiliser, conjointement à la double intégration, une estimation du mouvement rapide utilisant les seuls magnétomètres. Cette dernière technique a été décrite plus haut (diminuer la contribution du ou des accéléromètres au profit du ou des magnétomètres dans le calcul du ou des angles) . D'une part, la double intégration fournit en théorie le mouvement complet, mais elle est sujette à dérive . D'autre part, les magnétomètres fournissent une estimation partielle (excluant les rotations autour de l'axe du champ terrestre) mais non sujette à dérive. Une fusion possible consiste à estimer le mouvement à partir de la double intégration, à en déduire des mesures magnétiques estimées, et à utiliser la différence entre ces dernières et les mesures magnétiques réelles pour corriger le mouvement estimé par une technique de type "descente de gradient". De plus, on peut mettre en oeuvre un algorithme de fusion entre la méthode de double intégration et la méthode de mesure absolue, en passant de l'une à l'autre non pas de façon discontinue mais graduellement, en diminuant progressivement la contribution absolue des accéléromètres et en augmentant progressivement l'influence de la double intégration lorsque le mouvement accélère, et inversement en phase de décélération. Pour ce faire (diminution et augmentation progressives) , on procède par exemple de la façon exposée ci-après. On estime l'état du mobile (position, vitesse, accélération) à partir du dernier état connu et de la double intégration des accéléromètres de haute précision. On en déduit des estimations des données magnétométriques et accelerometriques du système de mesure absolue des angles (centrale d'attitude). On calcule une distance entre ces estimations et les mesures réelles. On corrige le mouvement estimé en appliquant une méthode de type "descente de gradient". La correction en question est paramétrée : elle est d'autant plus importante que le mouvement est lent, le critère étant par exemple le rapport de la norme du vecteur accélération à la norme de g, rapport qui a été mentionné plus haut. Ainsi, un mouvement très lent n'utilise pas les informations provenant de la double intégration, alors q'un mouvement très rapide les utilise exclusivement . L'invention présente tous les avantages de la technique qui est décrite dans le document [1] : elle peut être mise en œuvre de façon peu coûteuse, elle ne nécessite aucun équipement extérieur, tels que des sources magnétiques ou des caméras , et elle peut être mise en œuvre avec des algorithmes robustes. De plus, l'invention conduit à des mesures fiables même dans le cas de mouvements rapides . En outre, la présente invention peut être mise en œuvre avec une centrale d'attitude dont la précision angulaire est inférieure ou égale à 1° et avec des accéléromètres à au moins 10 bits (avantageusement de 14 à 16 bits) . On précise que le critère de « lenteur du mouvement », que l'on a mentionné plus haut, est fonction de la précision que l'on veut obtenir sur le mouvement . Un but de 1 ' invention étant de séparer l'accélération du solide de l'accélération de la pesanteur, tant que la norme de l'accélération du solide reste inférieure à — Il g II (peu différent de lm/s2) , le mouvement sera considéré comme lent et le procédé mènera à une précision acceptable. This example is schematically illustrated in FIG. 3. It is limited to the arm in the following but generalizable to the whole of the body. In FIG. 3, the references 26, 28, 30, 32 and 34 respectively represent the shoulder, the arm, the elbow, the forearm and the hand. The initial and final positions of the hand have the references 36 and 38 respectively. The movement of the hand, which performs a vertical translation of amplitude D, results in a rotation of angle around the elbow and possibly, depending on the amplitude of the translation, by another rotation around the shoulder. Instead of measuring D directly, we can therefore measure α. Knowing the length r of the forearm, we deduce the amplitude D of the translation. This technique according to the invention has the advantage of only being based on measurements absolute: it is therefore free from drifts. It should be noted that this technique uses, among other things, one or more accelerometers (attached to the forearm but not shown) to measure the angle α. This supposes that such sensors measure gravity and therefore make it possible to know their respective inclinations relative to the vertical. In the case of a rapid movement, the accelerometer (s) additionally measure the acceleration resulting from such a movement, so that the angle measurement is distorted. A technique known from document [1] partially solves this problem. It consists in reducing the contribution of the accelerometer (s) in favor of the magnetometer (s) in the calculation of the angle (s). But this technique is only partially effective and depends on the movement performed. A technique proposed in the present invention is capable of supplementing the previous one and is not limited by the type of movement allowed. This technique is explained below. As soon as a rapid movement is detected (it suffices to calculate the standard of the acceleration vector), the displacement of the sensor is calculated by double integration, taking as a starting point the state of the mobile solid (the arm in the example) at the start of the rapid phase, we simultaneously correct any possible drift, by merging the agnometric data, and as soon as the movement slows down, we return to absolute mode. This eliminates a possible drift which is likely to appear during the integration phase. The fusion of magnetometric data consists in using, in conjunction with double integration, an estimate of rapid movement using only magnetometers. The latter technique has been described above (reducing the contribution of the accelerometer (s) in favor of the magnetometer (s) in the calculation of the angle (s)). On the one hand, double integration provides in theory the complete movement, but it is subject to drift. On the other hand, magnetometers provide a partial estimate (excluding rotations around the axis of the earth's field) but not subject to drift. A possible fusion consists in estimating the movement from the double integration, in deducing therefrom estimated magnetic measurements, and in using the difference between the latter and the real magnetic measurements to correct the motion estimated by a technique of “gradient descent” type. ". In addition, a fusion algorithm can be implemented between the double integration method and the absolute measurement method, by passing from one to the other not discontinuously but gradually, gradually decreasing the absolute contribution of accelerometers and in gradually increasing the influence of double integration when the movement accelerates, and vice versa during the deceleration phase. To do this (gradual decrease and increase), for example, proceed as described below. The state of the mobile is estimated (position, speed, acceleration) from the last known state and the double integration of high precision accelerometers. We deduce estimates of the magnetometric and accelerometric data from the absolute angle measurement system (central attitude). A distance is calculated between these estimates and the actual measurements. The estimated movement is corrected by applying a "gradient descent" type method. The correction in question is parameterized: it is all the more important the slower the movement, the criterion being for example the ratio of the norm of the vector acceleration to the norm of g, ratio which was mentioned above. Thus, a very slow movement does not use the information coming from the double integration, whereas a very fast movement uses it exclusively. The invention has all the advantages of the technique which is described in document [1]: it can be implemented inexpensively, it does not require any external equipment, such as magnetic sources or cameras, and it can be implemented with robust algorithms. In addition, the invention leads to reliable measurements even in the case of rapid movements. In addition, the present invention can be implemented with an attitude unit whose angular precision is less than or equal to 1 ° and with accelerometers at least 10 bits (advantageously from 14 to 16 bits). It is specified that the criterion of "slowness of the movement", which we mentioned above, is a function of the precision that we want to obtain on the movement. An object of the invention being to separate the acceleration of the solid from the acceleration of gravity, as long as the standard of the acceleration of the solid remains below - Il g II (little different from lm / s 2 ), the movement will be considered slow and the process will lead to acceptable accuracy.

Claims

REVENDICATIONS 1. Procédé de mesure du mouvement d'un solide (2, 16, 32), procédé dans lequel on mesure au moins une première translation de ce solide, ce procédé comprenant une série d' étapes de mesure de l'accélération du solide et de double intégration des mesures ainsi effectuées pour obtenir des valeurs successives de la première translation, ce procédé étant caractérisé en ce qu'il comprend en outre une série d'étapes de mesure absolue d'au moins un deuxième degré de liberté du solide, ce deuxième degré de liberté étant une rotation, à l'aide d'au moins un capteur de rotation (6, 8) , et en ce que l'on transforme cette mesure de rotation en une mesure de translation et cette mesure de translation est utilisée pour recaler la première translation.CLAIMS 1. Method for measuring the movement of a solid (2, 16, 32), method in which at least a first translation of this solid is measured, this method comprising a series of steps for measuring the acceleration of the solid and of double integration of the measurements thus carried out to obtain successive values of the first translation, this method being characterized in that it further comprises a series of steps of absolute measurement of at least a second degree of freedom of the solid, this second degree of freedom being a rotation, using at least one rotation sensor (6, 8), and in that this measurement of rotation is transformed into a translation measurement and this translation measurement is used to reset the first translation. 2. Procédé selon la revendication 1, dans lequel on utilise la mesure du deuxième degré de liberté, obtenue à cette étape, en tant que condition initiale pour obtenir, par double intégration, la valeur de la première translation qui suit les valeurs précédemment obtenues de cette première translation. 2. Method according to claim 1, in which the measurement of the second degree of freedom, obtained at this stage, is used as an initial condition for obtaining, by double integration, the value of the first translation which follows the previously obtained values of this first translation. 3. Procédé selon l'une quelconque des revendications 1 et 2, dans lequel chaque étape de mesure absolue a lieu en même temps qu'une étape de mesure de l'accélération du solide (2, 16, 32). 3. Method according to any one of claims 1 and 2, wherein each absolute measurement step takes place at the same time as a measurement step of the acceleration of the solid (2, 16, 32). 4. Procédé selon l'une quelconque des revendications 1 à 3, dans lequel la transformation de la mesure de rotation en une mesure de translation utilise des modèles cinétiques du solide et/ou de son mouvement, permettant d'établir des relations entre la rotation et la translation.4. Method according to any one of claims 1 to 3, wherein the transformation of the measurement of rotation in a measurement of translation uses kinetic models of the solid and / or its movement, making it possible to establish relations between rotation and translation. 5. Procédé selon l'une quelconque des revendications 1 à 4, dans lequel le capteur de rotation est choisi parmi les accéléromètres (6) et les magnétomètres (8) .5. Method according to any one of claims 1 to 4, in which the rotation sensor is chosen from accelerometers (6) and magnetometers (8). 6. Procédé selon l'une quelconque des revendications 1 à 5, dans lequel la première translation est mesurée à l'aide d'un capteur de translation qui est aussi la capteur de rotation.6. Method according to any one of claims 1 to 5, wherein the first translation is measured using a translation sensor which is also the rotation sensor. 7. Procédé selon l'une quelconque des revendications 1 à 6, dans lequel on choisit un critère de lenteur du mouvement et, si le mouvement satisfait à ce critère après l'une des étapes de mesure du deuxième degré de liberté, on utilise la mesure du deuxième degré de liberté, obtenue à cette étape, pour recaler la première translation.7. Method according to any one of claims 1 to 6, in which a criterion of slowness of the movement is chosen and, if the movement satisfies this criterion after one of the steps of measuring the second degree of freedom, the measurement of the second degree of freedom, obtained at this stage, to readjust the first translation. 8. Procédé selon la revendication 7, dans lequel le critère de lenteur du mouvement est la situation d'une fonction de la norme de l'accélération du solide (2, 16, 32) en deçà d'un seuil prédéfini.8. The method of claim 7, wherein the criterion of slowness of the movement is the situation of a function of the standard of the acceleration of the solid (2, 16, 32) below a predefined threshold. 9. Procédé selon la revendication 8, dans lequel la fonction de la norme de l'accélération du solide (2, 16, 32) est cette norme elle-même. 9. The method of claim 8, wherein the function of the standard of the acceleration of the solid (2, 16, 32) is this standard itself.
EP04816578A 2003-12-22 2004-12-20 Method of sensing the motion of a solid, using an absolute measurement that is associated with a measurement calculated by double integration Ceased EP1714112A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
FR0351167A FR2864225B1 (en) 2003-12-22 2003-12-22 METHOD FOR MEASURING THE MOVEMENT OF A SOLID, USING AN ABSOLUTE MEASUREMENT ASSOCIATED WITH A DOUBLE INTEGRATION MEASUREMENT
PCT/FR2004/050729 WO2005064271A1 (en) 2003-12-22 2004-12-20 Method of sensing the motion of a solid, using an absolute measurement that is associated with a measurement calculated by double integration

Publications (1)

Publication Number Publication Date
EP1714112A1 true EP1714112A1 (en) 2006-10-25

Family

ID=34630618

Family Applications (1)

Application Number Title Priority Date Filing Date
EP04816578A Ceased EP1714112A1 (en) 2003-12-22 2004-12-20 Method of sensing the motion of a solid, using an absolute measurement that is associated with a measurement calculated by double integration

Country Status (5)

Country Link
US (1) US7460975B2 (en)
EP (1) EP1714112A1 (en)
JP (1) JP2007515637A (en)
FR (1) FR2864225B1 (en)
WO (1) WO2005064271A1 (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2864225B1 (en) 2003-12-22 2006-07-21 Commissariat Energie Atomique METHOD FOR MEASURING THE MOVEMENT OF A SOLID, USING AN ABSOLUTE MEASUREMENT ASSOCIATED WITH A DOUBLE INTEGRATION MEASUREMENT
FR2895500B1 (en) * 2005-12-23 2008-03-28 Commissariat Energie Atomique METHOD OF ESTIMATING MOVEMENT OF A SOLID
EP2271946B8 (en) * 2008-04-30 2019-06-26 Movea S.A. Device for detecting a percussion event, and associated mobile system
GB2574074B (en) 2018-07-27 2020-05-20 Mclaren Applied Tech Ltd Time synchronisation
GB2588236B (en) 2019-10-18 2024-03-20 Mclaren Applied Ltd Gyroscope bias estimation

Family Cites Families (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4119212A (en) 1977-07-18 1978-10-10 Western Electric Company, Inc. Monitoring the location of a robot hand
JPS63315909A (en) * 1987-06-19 1988-12-23 Fujitsu Ltd inertial navigation device
JP2894872B2 (en) * 1991-06-28 1999-05-24 日本電気ホームエレクトロニクス株式会社 Navigation device
JPH0755513A (en) * 1993-08-19 1995-03-03 Canon Inc Position managing device
SE501867C2 (en) * 1993-11-15 1995-06-12 Asea Brown Boveri Method and system for calibrating an industrial robot using a spherical calibration body
US5615132A (en) 1994-01-21 1997-03-25 Crossbow Technology, Inc. Method and apparatus for determining position and orientation of a moveable object using accelerometers
JPH07295736A (en) * 1994-04-25 1995-11-10 Sony Corp Three-dimensional position detector
US5645077A (en) * 1994-06-16 1997-07-08 Massachusetts Institute Of Technology Inertial orientation tracker apparatus having automatic drift compensation for tracking human head and other similarly sized body
JPH08285621A (en) * 1995-04-14 1996-11-01 Omron Corp Navigation device
JPH09189564A (en) * 1996-01-11 1997-07-22 Matsushita Electric Ind Co Ltd Traveling body position speed calculating device
JPH09257461A (en) * 1996-03-18 1997-10-03 Ricoh Co Ltd Three-dimensional coordinate measuring apparatus
US20030047002A1 (en) * 1998-10-28 2003-03-13 Steven W. Arms Mems based angular accelerometer
FR2826447B1 (en) * 2001-06-26 2003-09-19 Sagem HYBRID INERTIAL NAVIGATION METHOD AND DEVICE
US6728632B2 (en) * 2001-08-30 2004-04-27 Ericsson Inc. Navigation devices, systems, and methods for determining location, position, and/or orientation information based on movement data generated by a movement detector
ATE442573T1 (en) * 2001-11-13 2009-09-15 Nokia Corp METHOD, DEVICE AND SYSTEM FOR CALIBRATION OF ANGLE RATE MEASUREMENT SENSORS
FR2838185B1 (en) * 2002-04-05 2004-08-06 Commissariat Energie Atomique DEVICE FOR CAPTURING ROTATIONAL MOVEMENTS OF A SOLID
FR2847689B1 (en) 2002-11-27 2005-01-21 Commissariat Energie Atomique METHOD AND DEVICE FOR CAPTURING THE MOVEMENT OF A SOLID USING AT LEAST ONE CAMERA AND AN ANGULAR SENSOR
FR2860700B1 (en) 2003-10-10 2005-12-09 Commissariat Energie Atomique CROWN CONTROL DEVICE
FR2864225B1 (en) 2003-12-22 2006-07-21 Commissariat Energie Atomique METHOD FOR MEASURING THE MOVEMENT OF A SOLID, USING AN ABSOLUTE MEASUREMENT ASSOCIATED WITH A DOUBLE INTEGRATION MEASUREMENT
FR2897680B1 (en) * 2006-02-17 2008-12-05 Commissariat Energie Atomique MOTION CAPTURE DEVICE AND ASSOCIATED METHOD

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See references of WO2005064271A1 *

Also Published As

Publication number Publication date
WO2005064271A1 (en) 2005-07-14
FR2864225A1 (en) 2005-06-24
FR2864225B1 (en) 2006-07-21
JP2007515637A (en) 2007-06-14
US7460975B2 (en) 2008-12-02
US20070163343A1 (en) 2007-07-19

Similar Documents

Publication Publication Date Title
EP1984696B1 (en) Motion capture device and associated method
EP1985233B1 (en) Method and device for detecting a substantially invariant axis of rotation
EP2400460B1 (en) Method for assessing the horizontal speed of a drone, particularly of a drone capable of hovering on automatic pilot
EP3658921B1 (en) Method for calibrating a magnetometer
FR2916069A1 (en) PROCESSING METHOD FOR MOTION CAPTURE OF ARTICULATED STRUCTURE
WO2004051391A2 (en) Method and device for capturing the movement of a solid using at least one camera and one angular sensor
WO2003085357A2 (en) Device for rotational motion capture of a solid
WO2012168357A1 (en) Simplified method for estimating the orientation of an object, and attitude sensor implementing such a method
WO2011144408A1 (en) Method and system for fusing data arising from image sensors and from motion or position sensors
EP2710334B1 (en) Method of calibrating an inertial assembly comprising a dynamic phase between two static phases
EP2938965B1 (en) Method for comparing two inertial units integral with a same carrier
EP2313739A1 (en) Method for the improved estimation of an object orientation and attitude control unit implementing said method
FR2987913A1 (en) IMPROVED MOTION CAPTURE SYSTEM OF ARTICULATED STRUCTURE
EP3427088B1 (en) Method for detecting an anomaly in the context of using a magnetic positioning device
EP1721573B1 (en) Method for estimating the phase of movement of an object
EP1969314A2 (en) Method for estimating movement of a solid
WO2005064271A1 (en) Method of sensing the motion of a solid, using an absolute measurement that is associated with a measurement calculated by double integration
FR2951535A1 (en) METHOD OF DETECTING PARASITE MOVEMENTS DURING THE ALIGNMENT OF AN INERTIAL POWER PLANT
FR3118493A1 (en) METHOD AND DEVICE FOR CONTROLLING THE POSITIONING DETERMINATION OF A WEARED INFORMATION DISPLAY DEVICE
EP2689211B1 (en) Method for updating a value of orientation with respect to north or for improving the initialization of such a value in an apparatus comprising an image sensor
EP3211370A1 (en) Method for filtering signals from a sensor assembly comprising at least one sensor for measuring a vector physical field that is substantially constant in time and space in a frame of reference
WO2018011498A1 (en) Method and system for locating and reconstructing in real time the posture of a moving object using embedded sensors
EP4077016B1 (en) System and method for providing visual assistance to an individual suffering from motion sickness
WO2024033583A1 (en) Method and device for processing a sequence of images for identifying tracked objects in the sequence of images
FR3041769A1 (en) GEOLOCATION PROCESS

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

17P Request for examination filed

Effective date: 20060529

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LI LT LU MC NL PL PT RO SE SI SK TR

RAP1 Party data changed (applicant data changed or rights of an application transferred)

Owner name: COMMISSARIAT A L'ENERGIE ATOMIQUE

DAX Request for extension of the european patent (deleted)
RAP1 Party data changed (applicant data changed or rights of an application transferred)

Owner name: COMMISSARIAT A L'ENERGIE ATOMIQUE ET AUX ENERGIES

17Q First examination report despatched

Effective date: 20110601

REG Reference to a national code

Ref country code: DE

Ref legal event code: R003

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION HAS BEEN REFUSED

18R Application refused

Effective date: 20131108