Method for estimating a trajectory of a point of interest

Abstract

This method for estimating a trajectory, over a displacement interval, of a point of interest belonging to a limb of an individual comprises estimating (1320) the evolution, during the displacement interval, of a speed of a measurement point integral with the point of interest, estimating (1310) the evolution of an orientation of the measurement point during the displacement interval, and estimating the trajectory of the point of interest during the displacement interval from the estimated speed and orientation of the measurement point.

Description

[0001]DESCRIPTION TITLE: METHOD FOR ESTIMATING A TRAJECTORY OF A POINT OF INTEREST FIELD OF THE INVENTION The present invention relates to the field of estimating the evolution of the linear speed of a point from measurements taken using an inertial device. The invention also relates to estimating the trajectory of a point of interest belonging to a limb of an individual from measurements taken using an inertial device, and the use of such an estimation to evaluate and analyze an effort exerted by said limb, in particular when this limb is an upper limb. TECHNOLOGICAL BACKGROUND For a person, moving their upper limbs depends on the use and coordination of several muscles. The analysis of an individual's movements can therefore be used to characterize their state of physical fitness,and in particular the muscular strength that the latter is capable of developing. This muscular strength is likely to vary depending on many factors such as the individual's fitness, age, physical training, or taking a treatment affecting their muscular efficiency. The evolution of an individual's movements over time may be representative of the evolution of their physical fitness, particularly in the case of an individual suffering from a neuromuscular or neurodegenerative disease. These diseases, such as Duchenne Muscular Dystrophy (DMD), result in a reduction in the individual's muscular capacities as the disease progresses. This leads to a loss of motor skills in the affected patients. WO 2019 / 243609 discloses a method for analyzing the stride of a walking pedestrian, intended to characterize the evolution of their fitness,from measurements taken using an inertial device attached to the pedestrian's ankle. Such a minimally invasive method finds particular application in monitoring neuromuscular or neurodegenerative diseases. However, it is only suitable for monitoring such diseases as long as the individual retains sufficient motor function in their lower limbs. When the individual is no longer able to walk, typically because the disease is at an advanced stage,An alternative method is needed to enable monitoring of the disease. A consensus has thus emerged in the scientific community on the need to be able to accurately quantify the motor function of the upper limbs of individuals suffering from neuromuscular or neurodegenerative diseases. This consensus has led to the emergence of several methods for assessing this motor function. The vast majority of these assessment methods consist of traditional methods based on functional scales assessed by professional clinicians in controlled environments. The individual is typically invited to go to a hospital where they are asked to perform certain movements and, depending on the success or failure of these movements, the clinician assigns a motor score. These methods include, for example: - the Brooke scale (Brooke,MH et al. “Clinical investigation in Duchenne dystrophy: 2. Determination of the “power” of therapeutic trials based on the natural history.” Muscle & nerve vol. 6.2 (1983): 91-103. doi:10.10002 / mus.880060204), - the Revised Upper Limb Module (RULM) (Mazzone, Elena S et al. “Revised upper limb module for spinal muscular atrophy: Development of a new module.” Muscle & nerve vol. 55,6 (2017): 869-874. doi:10.10002 / mus.25430), and - the Performance of Upper Limb Module (PULM) (Mayhew, Anna G et al. “Performance of Upper Limb module for Duchenne muscular dystrophy.” Developmental medicine and child neurology vol. 62,5 (2020): 633-639. doi:10.1111 / dmcn.14361). However, these methods have several drawbacks: they are restrictive for the patient,by requiring them to travel to a hospital, and their results may vary depending on the patient's fitness level on the day of the assessment and the assessor. Recently, other methods have emerged, based on the use of information technology. Most use the analysis of data recorded in a controlled environment, typically in a hospital setting, using an optical capture system under the supervision of a healthcare professional, as described for example in Han, JJ et al. “Reachable workspace and performance of upper limb (PUL) in Duchenne muscular dystrophy.” Muscle & nerve vol. 53 (2016): 545-554. doi:10.10002 / mus.24894. However, these methods suffer from most of the drawbacks already identified for traditional methods, notably because they are restrictive for the patient and produce results that vary depending on the patient's fitness level on the day of the assessment. To solve these problems,Solutions have been sought to enable the assessment of upper limb motor function over long acquisition times, in the daily lives of patients, most often by means of measurements carried out by inertial devices attached to the upper limbs. These inertial devices must meet particularly demanding constraints of weight, size and autonomy to allow their portability, they are composed of MEMS type sensors. However, such sensors have a significant time drift which greatly complicates the monitoring of movements using the measurements returned by these sensors. To circumvent this difficulty, it was proposed in WO 2017 / 129890 to assess the motor function of the upper limbs only for very specific movements, when the elbow is placed motionless on a surface and the movement of the wrist, to which the inertial device is attached,follows an arc centered on the elbow. However, this solution is not entirely satisfactory, as many upper limb movements are not taken into account in this assessment of motor function. It has also been proposed in El-Gohary, Mahmoud, and James McNames. “Shoulder and elbow joint angle tracking with inertial sensors.” IEEE transactions on bio-medical engineering vol. 59,9 (2012): 2635-41. doi:10.11009 / TBME.2012.2208750 to combine a kinematic model of the upper limb with measurements acquired by two inertial devices positioned one on the arm, the other on the forearm, to compensate for sensor drift. This solution, which uses several inertial devices, is however not very suitable for daily use, as it is too invasive. Finally, it has been proposed in Villeneuve, Emma et al. “Reconstruction of Angular Kinematics From Wrist-Worn Inertial Sensor Data for Smart Home Healthcare.” IEEE Access,5 (2017): 2351-2363. doi:10.11009 / access.2016.2640559 to estimate the kinematics of a person's wrist from measurements provided by simple accelerometers worn on the wrist, through the use of a particle filter and the imposition of constraints related to compliance with a biomechanical model of the person's upper limb. However, while this method makes it possible to estimate wrist kinematics without using a gyrometer, it does not solve the problem of sensor time drift. Other difficulties encountered in assessing upper limb motor function using inertial devices are that upper limb movements are much more varied and difficult to recognize than lower limb movements and that, unlike the latter, upper limb movements are often linked not only to their own muscular activity,but also to the muscular activity of other parts of the body, in particular the trunk or the lower limbs. DISCLOSURE OF THE INVENTION One objective of the invention is to enable the monitoring of the physical fitness of an individual over time. Another objective is to enable a good estimation, in an uncontrolled environment and over a long period, of the physical efforts made by an upper limb of an individual for various movements. Other objectives are to minimize the invasive nature of the devices used to carry out this estimation, and to improve the quality of the estimations deduced from the measurements of an inertial device. To this end, the invention relates, according to a first aspect, to a method for estimating the evolution of a linear speed of a measurement point during a movement interval,the method being implemented by a data processing unit and comprising the following steps: - determining an initial speed and a final speed of the measurement point respectively at the start and at the end of the movement interval, - estimating the evolution of a speed of the measurement point during the movement interval, said estimation comprising: o the downstream integration of movement parameters of the measurement point acquired by an inertial device during the movement interval, taking as the initial value of the speed the value of the initial speed, o the upstream integration of the movement parameters, taking as the initial value of the speed the value of the final speed, and o the calculation of the estimated speed by merging the upstream and downstream integrations. The invention also relates, according to a second aspect, to a method for estimating a trajectory, over a movement interval,of a point of interest belonging to a member of an individual, the method being implemented by a data processing unit and comprising the following steps: - estimation of the evolution, during the movement interval, of a speed of a measurement point integral with the point of interest, - estimation of the evolution of an orientation of the measurement point during the movement interval, and - estimation of the trajectory of the point of interest during the movement interval from the estimated speed and orientation of the measurement point. According to particular embodiments of the invention, the trajectory estimation method also has one or more of the following characteristics,taken in isolation or in any technically possible combination(s): - the individual is a human being; - the limb is an upper limb of the individual; - the point of interest is a point on a forearm of said upper limb; - the limb is a lower limb of the individual; - the point of interest is a point on a leg of said lower limb; - the evolution of the speed of the measurement point during the movement interval is estimated by means of a method according to the first aspect; - the estimation of the trajectory comprises the following steps: o calculation of several candidate trajectories by applying the estimated speed and orientation to different pairs of starting and finishing configurations of a biomechanical model of the limb, o evaluation of a compatibility of each candidate trajectory with the biomechanical model,and o selection of the candidate trajectory with the best compatibility; - the point of interest is a point on a forearm belonging to an upper limb of the individual, each starting configuration of the biomechanical model of the limb being defined by a pair of starting angles formed by a first starting angle between the forearm and the vertical and a second starting angle between the arm and the vertical and each arrival configuration of the biomechanical model of the limb being defined by a pair of arrival angles formed by a first arrival angle between the forearm and the vertical and a second arrival angle between the arm and the vertical; - for each starting or arrival configuration of the biomechanical model of the limb: o the first starting angle, respectively arrival angle, is deduced from the orientation of the measurement point at the beginning, respectively at the end, of the displacement interval, and o the second starting angle, respectively arrival angle,is less than the first departure angle, respectively the arrival angle; - there exists, for each value of the second departure angle less than the first departure angle and each value of the second arrival angle included in a predetermined interval, a pair of departure and arrival configurations for which a candidate trajectory is calculated; - the predetermined interval is constituted by an interval between 0° and the value of the first arrival angle, or by an angular interval around an angle β, t verifying the following relation: ^(cos in which: o L is the arm length of the upper limb, o βi is the value of the second starting angle in the starting configuration belonging to the same pair as the arrival configuration, o ti is the start time of the displacement interval, ot f is the end time of the displacement interval, o ^ ^^^^ ( ^^ ^ ^^) is the velocity vector of the measurement point at each instant t of the displacement interval, ol is the distance between the elbow of the upper limb and the measurement point, o ^ ^ ^^^ ^ is a unit vector giving the orientation of a principal axis of the forearm at the end of the displacement interval, o ^ ^ ^^ ^ is a unit vector giving the orientation of the main axis of the forearm at the start of the displacement interval, where ^ is a unit vector giving the vertical orientation, the vectors ^ ^^^^ ( ^^ ^ ^^ ) , ^ ^^^^ ^ , ^ ^^^ ^and ^ being expressed in the same inertial frame; - the angular interval extends over less than 20°, preferably over less than 15°, for example over substantially 10°; and - the calculation of each candidate trajectory comprises the following steps: o downstream integration of the estimated speed, taking as initial position the position of the measurement point in the starting configuration, o upstream integration of the estimated speed, taking as initial position the position of the measurement point in the arrival configuration, and o calculation of the candidate trajectory by merging the upstream and downstream integrations; The invention also relates, according to a third aspect, to a method for evaluating an effort exerted by a limb of an individual,the method being implemented by a data processing unit and comprising the following steps: - estimation of a trajectory of a point of interest belonging to the member over at least one primary displacement interval by means of a method according to the second aspect, - evaluation, from the estimated trajectory and speed evolution, of a primary force exerted by the member during the or each primary displacement interval. According to particular embodiments of the invention, the force evaluation method also has one or more of the following characteristics, taken in isolation or in any technically possible combination(s): - the method comprises the following additional steps: o deduction, from the displacement parameters acquired during the primary displacement interval,of a rotational component of a primary kinetic energy acquired by at least one portion of the member during the primary displacement interval, the primary force being a function of said rotational component of the primary kinetic energy, o determination of a relationship between the primary force and the rotational component of the primary kinetic energy, o estimation of the evolution of an orientation of a measurement point integral with the point of interest during at least one secondary displacement interval, o deduction, from displacement parameters acquired by the inertial device during a secondary displacement interval, of a rotational component of a secondary kinetic energy acquired by said at least one portion of the member during the secondary displacement interval, o evaluation, from the determined relationship and the rotational component of the secondary kinetic energy,of a secondary effort exerted by the limb during the or each secondary movement interval; - the method comprises, for each of the primary and secondary movement intervals: o the estimation of the evolution, during the primary or secondary movement interval, of a speed of a measurement point integral with the point of interest, o the calculation of several candidate trajectories by applying the estimated speed and orientation to different pairs of starting and finishing configurations of a biomechanical model of the limb, o the evaluation of a compatibility of each candidate trajectory with the biomechanical model, and o the verification, for each primary or secondary movement interval, of the acceptable nature of the compatibility of at least one candidate trajectory,the trajectory of the point of interest being estimated only when at least one candidate trajectory has acceptable compatibility; and - the evolution of the speed of the measurement point during the primary or secondary displacement interval is estimated by means of a method according to the first aspect. The invention also relates, according to a fourth aspect, to a method for analyzing an effort exerted by a limb of an individual, the method being implemented by a data processing unit and comprising the following steps: - a) estimation, for several intervals of displacement of a point of interest of the limb occurring during a predetermined period, of an effort exerted by said limb by means of a method according to the third aspect, - b) determination of the value of at least one statistical quantity, calculated on the set formed by the estimated efforts, representative of a state of fitness of the individual. According to a particular embodiment of the invention,the effort analysis method also has the following characteristic: - the method comprises repeating steps a) and b) for several predetermined periods and observing the evolution of the or each statistical quantity across the different predetermined periods. The invention also relates, according to a fifth aspect, to a method for monitoring the fitness state of an individual, comprising the analysis of an effort exerted by a limb of the individual by means of an analysis method according to the fourth aspect and the deduction of a fitness state of the individual from said analysis. According to a particular embodiment of the invention, the monitoring method also has the following characteristic: - the method comprises administering a treatment to the individual, and evaluating an effectiveness of the treatment from the deduced fitness state. The invention finally relates, according to a fifth, respectively,a sixth and a seventh aspect of the invention: - equipment comprising an inertial device for acquiring displacement parameters of a measurement point during at least one displacement interval, and a data processing unit configured to estimate, from the displacement parameters acquired by the inertial device, by implementing a method according to any one of the first, second and third aspects, at least one of: the evolution of a speed of the measurement point during the displacement interval, a trajectory, over the displacement interval, of a point of interest integral with the measurement point, and a force exerted by a limb of an individual to which the point of interest belongs; - a computer program product comprising code instructions for implementing a method according to any one of the first, second, third,fourth and fifth aspects when said program is executed by a computer; and - a storage means readable by computer equipment on which is stored a computer program product comprising code instructions for implementing a method according to any one of the first, second and third aspects when said program is executed by a computer. BRIEF DESCRIPTION OF THE FIGURES Other characteristics and advantages of the invention will appear on reading the following description, given solely by way of example and with reference to the appended drawings, in which: - Figure 1 is a diagram of equipment according to an exemplary embodiment of the invention, - Figure 2 is a diagram of a measuring device of the equipment of Figure 1, - Figure 3 is a diagram of a motion sensor of the equipment of Figure 1,- Figure 4 is a diagram illustrating a method of force analysis implemented by the equipment of Figure 1, - Figure 5 is a diagram detailing a step of the method of Figure 4, - Figure 6 is a diagram detailing a first sub-step of the step of Figure 5, - Figure 7 is a diagram detailing a second sub-step of the step of Figure 5, - Figure 8 is a diagram detailing a third sub-step of the step of Figure 5, - Figure 9 is a diagram of a biomechanical model of an upper limb used for implementing the method of Figure 4, and - Figure 10 is a diagram showing a pair of starting and ending configurations of the biomechanical model of Figure 9. DETAILED DESCRIPTION OF AN EXAMPLE OF EMBODIMENT The equipment 10 shown in Figure 1 is intended for the evaluation and analysis of the forces exerted by the muscles of an upper limb 12 of a person 14. By “person”,here and hereinafter, a human being is understood. For this purpose, the equipment 10 comprises a measuring device 20 attached to a forearm 22 of the upper limb 12. In the example shown, it also comprises a motion sensor 24 intended to capture the movement of at least one lower limb of the person 14. It is recalled here that, anatomically, an upper limb such as the upper limb 12 is made up of three segments: the arm 25, articulated to the trunk 26 by the shoulder 27, the forearm 22, articulated to the arm 25 by the elbow 28, and the hand 29, articulated to the forearm 22 by the wrist 30. Here and hereinafter, it is in this sense that these terms are used, the term "arm" being in particular used to designate only the segment of the upper limb between the elbow and the shoulder and not, as in common language, the entire upper limb. The measuring device 20 is secured to the forearm 22 of the upper limb 12,that is to say that it presents, in the terrestrial reference frame, a movement substantially identical to that of the forearm 12. It is in particular placed between the elbow 28 (excluded) and the wrist 30 (included), for example, as shown, substantially on the wrist 30. With reference to Figure 2, the measuring device 20 comprises a support 31 and an attachment member 32 for attaching the support 31 to the forearm 22. It also comprises an inertial device 33 and a data processing unit 34 mounted on the support 31. In the example shown, it also comprises a communication system 36, typically a wireless communication system, mounted on the support 31 for communication of the measuring device 20 with the second measuring device 24 and / or an external device such as a mobile terminal (not shown), for example a multifunction mobile,or a remote server (not shown). Optionally, it also comprises a storage module 38 mounted on the support 31. The support 31 is typically constituted by a housing. The attachment member 32 is here constituted by a bracelet, for example with a self-gripping band, adapted to encircle the forearm 22 and allow the integral connection. Alternatively (not shown), the attachment member 32 is constituted by any element allowing an integral connection with the forearm 22. The inertial device 33 comprises a gyrometer 40 for measuring an angular velocity of the measuring device 20 according to a system of three orthogonal axes defining a mobile reference frame (Rm) integral with the measuring device 20, that is to say measuring the three components of an angular velocity vector in said mobile reference frame (Rm). It is thus understood that the gyrometer 40 is typically made up of a set of three gyrometers each associated with one of the three axes,in particular in tri-axis (i.e. each one capable of measuring one of the three components of the angular velocity vector). Preferably, the mobile reference frame (Rm) is chosen so that one of its axes is collinear with the forearm 22 when the measuring device 20 is worn. Here and in the following, the mobile reference frame (Rm) is defined as being a direct orthogonal reference frame formed of a triplet of axes, shown in Figure 2, comprising: - an axis U collinear with the forearm 22 and oriented towards the hand 29, - an axis V orthogonal to the axis to the U axis, and - an axis W orthogonal to the U and V axes. The inertial device 33 also comprises an accelerometer 42 for measuring an acceleration of the measuring device 20 according to a system of three orthogonal axes defining a mobile reference frame integral with the measuring device 20, which is advantageously the same as the mobile reference frame (R, m) of the gyrometer 40. In other words, the accelerometer is capable of measuring the three components of an acceleration vector in said moving frame. It is thus understood that the accelerometer 42 is typically made up of a set of three accelerometers each associated with one of the three axes, in particular in tri-axis (i.e. each capable of measuring one of the three components of the acceleration vector). These accelerometers are sensitive to external forces, including gravitational forces, applied to the inertial device 33, and make it possible to measure a specific acceleration. The data processing unit 34 is configured to deduce from the measurements of the inertial device 33 an orientation, a linear speed and a position of a measurement point integral with the measuring device 20, for example constituted by the origin of the moving frame (R m) of the gyrometer 40 and the accelerometer 42. It is also configured to estimate a linear speed and a position of the center of gravity of the forearm 22. It is further configured to evaluate and analyze the efforts of the muscles of the upper limb 12. For this purpose, the data processing unit 34 is, in the example shown, constituted by a programmable machine, such as a DSP ("Digital Signal Processor" in English) or a microcontroller. It comprises a processor or CPU ("Central Processing Unit" in English) 44 and a memory 46 of the RAM ("Random Access Memory" in English) and / or ROM ("Read Only Memory" in English) type. The processor 44 is configured to execute instructions loaded into the memory 46. When the measuring device 20 is powered on, the processor 44 is capable of reading instructions from the memory 46 and executing them.These instructions form a computer program causing the implementation, by the processor 44, of certain steps of a method 1000 (Figure 4) which will be detailed below. Alternatively (not shown), the data processing unit 34 is constituted by a machine or a dedicated component, such as an FPGA (“Field-Programmable Gate Array”) or an ASIC (“Application-Specific Integrated Circuit”). The data processing unit 34 further comprises a buffer memory 48 for the temporary storage of information necessary for the implementation of the method 1000. The communication system 36 is configured to implement short-range wireless communication, for example Bluetooth or Wi-Fi (in particular in the example shown with the motion sensor 24) and / or to connect to a mobile network (typically UMTS / LTE / 5G) for long-distance communication.Alternatively (not shown), the communication system 36 is for example a wired connection (typically USB) for transferring the data from the storage module 38 to another storage module, typically a server (not shown). For example, the communication system 36 is configured to receive from the motion sensor 24 movement data of at least one lower limb of the person 14 and / or to transmit the analysis data to an external device for presentation to a human operator, typically for display on a screen. Optionally, the measuring device 20 also comprises a network of magnetometers (not shown) linked to the support 31, that is to say that they each have a movement substantially identical to that of the support 31 in the terrestrial reference frame, and spatially spaced from each other. Each magnetometer is a tri-axis magnetometer capable of measuring a magnetic field along three axes.For this purpose, each magnetometer is typically constituted by three single-axis magnetometers (not shown) oriented along axes substantially perpendicular to each other. These axes are preferably the same as those of the system of three orthogonal axes of the gyrometer 40 and / or the accelerometer 42. The network of magnetometers is suitable, due to its particular geometry, for allowing the determination, at each measurement instant of the magnetometers, of a spatial gradient of the measured magnetic field, in particular of the coefficients of this gradient along each of the axes U, V, W of the moving reference frame (Rm) linked to the support 31. Each coefficient of said gradient is for example determined by a method using the vector measurements of the magnetic field carried out by the magnetometers, associated with optimization methods of the least square or median filter type or associated with the intrinsic properties of the magnetic field described by Maxwell's equations.However, any other conventional method suitable for calculating the coefficients of the spatial gradient of the magnetic field is suitable. The data processing unit 34 is then typically configured to determine the linear speed of the measurement point, over certain displacement phases, by implementing the method described in EP 2541199. In the example described above, the processing unit 34 of the measuring device 20 is mounted on the support 31. Alternatively (not shown), at least part of the data processing unit 34 is remote, for example in a mobile terminal (not shown) and / or in a remote server (not shown). In other words, at least part of the steps of the method 1000 is performed by a mobile terminal and / or a remote server.The communication system 36 is then configured to send to the mobile terminal and / or to the remote server the data from the inertial device 33 and, where appropriate, from the magnetometer network of the first measuring device 20. Returning to Figure 1, the motion sensor 24 is here secured to a lower limb 50 of the person 14. It is recalled here that, anatomically, a lower limb such as the lower limb 50 is made up of three segments: the thigh 51, articulated to the trunk 26 by the hip 52, the leg 53, articulated to the thigh 51 by the knee 54, and the foot 55, articulated to the leg 53 by the ankle 56. Here and in the following, it is in this sense that these terms are used, the term "leg" being in particular used to designate only the segment of the upper limb between the knee and the ankle and not, as in common language, the entire lower limb.The motion sensor 24 is in particular secured to the leg 53 of the lower limb 50, that is to say that it presents, in the terrestrial reference frame, a movement substantially identical to that of the leg 53. It is in particular placed between the knee 54 (excluded) and the ankle 56 (included), for example, as shown, substantially on the ankle 56. Alternatively, the motion sensor 24 is fixed to a wheelchair 58 on which the person 14 is seated. With reference to Figure 3, the motion sensor 24 here comprises a support 61 and an attachment member 62 for attaching the support 61 to the leg 53. It also comprises a gyrometer 63 and a data processing unit 64 mounted on the support 61.In the example shown, it also comprises a communication system 66, typically a wireless communication system, mounted on the support 61 for communication of the motion sensor 24 with the measuring device 20 and / or an external device such as a mobile terminal (not shown), for example a multifunction mobile, or a remote server (not shown). Optionally, it also comprises a storage module 68 mounted on the support 61. The support 61 is typically constituted by a housing. The attachment member 62 is here constituted by a bracelet, for example with a hook-and-loop strip, adapted to enclose the leg 53 and allow the integral connection. As a variant (not shown), the attachment member 62 is constituted by any element allowing an integral connection with the leg 53.The gyrometer 63 is capable of measuring an angular velocity of the motion sensor 24 according to a system of three orthogonal axes defining a moving frame integral with the motion sensor 24, that is to say measuring the three components of an angular velocity vector in said moving frame. It is thus understood that the gyrometer 63 is typically constituted by a set of three gyrometers each associated with one of the three axes, in particular in tri-axis (i.e. each capable of measuring one of the three components of the angular velocity vector). The data processing unit 64 is configured to deduce from the measurements of the gyrometer 63 the movement phases and the immobility phases of the motion sensor 24.For this purpose, the data processing unit 64 is typically configured to compare the norm of the angular velocity measured by the gyrometer 63 with a threshold, and to define the phases during which this norm is lower than the threshold as being phases of immobility and the phases during which this norm is lower than the threshold as being phases of movement. The data processing unit 64 is, in the example shown, constituted by a programmable machine, such as a DSP (Digital Signal Processor) or a microcontroller. It comprises a processor or CPU (Central Processing Unit) 74 and a memory 76 of the RAM (Random Access Memory) and / or ROM (Read Only Memory) type. The processor 74 is configured to execute instructions loaded into the memory 76.When the motion sensor 24 is powered on, the processor 74 is capable of reading instructions from the memory 76 and executing them. These instructions form a computer program causing the processor 74 to implement a method for detecting a movement of the motion sensor 24. Alternatively (not shown), the data processing unit 64 is constituted by a machine or a dedicated component, such as an FPGA (Field-Programmable Gate Array) or an ASIC (Application-Specific Integrated Circuit). The data processing unit 64 further comprises a buffer memory 78 for the temporary storage of information necessary for detecting the movement of the motion sensor 24.The communication system 66 is configured to implement short-range wireless communication, for example Bluetooth or Wi-Fi (in particular in the example shown with the measuring device 20) and / or to connect to a mobile network (typically UMTS / LTE / 5G) for long-distance communication. Alternatively (not shown), the communication system 66 is, for example, a wired connection (typically USB) for transferring the data from the storage module 68 to another storage module, typically a server (not shown). For example, the communication system 66 is configured to transmit the start and end times of the immobility phases to the measuring device 20 and / or to an external device, typically a mobile telephone or a server. A method 1000 implemented by the equipment 10 will now be described, with reference to Figures 4 to 8.As seen in Figure 4, the method 1000 begins with a step 1002 of acquiring displacement parameters of the measurement point. During this step, the inertial device 33 acquires the acceleration ^(t. k ) and the rotation speed ^^ (t k ) of the measuring point at each of a plurality of sampling times tk. It will be noted that the acceleration ^(tk) and the rotation speed ^^ (tk) are here vectors each having three components along each of the axes U, V, W of the moving frame (Rm) attached to the support 31 of the measuring device 20. The sampling times t kare spaced from each other by a time dt which is very small compared to the characteristic time of the movements of the person 14, this time dt being for example substantially equal to 10 ms. Step 1002 follows a start-up of the measuring device 20 and typically lasts until the measuring device 20 is switched off. The method 1000 also comprises, following step 1002 or, more precisely, following the acquisition of acceleration vectors ^(t k ) and rotation speed ^^ (t k ) for enough sampling times t k , a step 1004 of determining an orientation of the measurement point relative to the forearm 22. This step makes it possible to know the orientation of the mobile reference point (R m) attached to the support 31 of the measuring device 20 relative to the forearm 22. Indeed, if the orientation of said mobile reference mark (Rm) relative to the support 31 is known, two positions of the support 31 relative to the forearm 22 are possible, depending on the direction in which the attachment member 32 is mounted on the forearm 22: a position in which the axis U of the mobile reference mark (Rm) collinear with the forearm 22 points towards the hand 29 and a position in which said axis points towards the elbow 28. For example, step 1004 comprises the detection, by the processing unit 34, of phases of immobility of the measuring device 20, followed by the measurement of the acceleration during said phases of immobility. The detection of immobility phases is typically carried out by comparing the rotation speed norm ^^ (tk) with a threshold, an immobility phase being detected when, for several sampling instants t ksuccessive, the standard of the rotation speed ^^ (tk) is lower than said threshold. The measurement of the acceleration during these phases of immobility gives the orientation of the gravitational acceleration in the moving frame (Rm), from which the processing unit 34 deduces the orientation of the measurement point relative to the forearm 22. If it results from this deduction that the axis U points towards the hand 29, then the moving frame (R m) is maintained. If, on the contrary, the U axis points towards the elbow 28, then the mobile reference frame (Rm) is reoriented (it is typically rotated 180° around its V axis or its W axis) so that the U axis points towards the hand 29. In parallel with steps 1002 and 1004, the method 1000 also comprises a step 1006 of detecting phases of immobility of the lower limbs of the person 14, followed by a step 1008 of transmitting the start and end times of these phases of immobility to the measuring device 20. During step 1006, the processing unit 64 of the motion sensor 24 typically compares the standard of the rotation speed measured by the gyrometer 63 with a threshold, and determines the phases of immobility as being the periods during which said standard is lower than the threshold.Then, during step 1008, the motion sensor 24 transmits to the measuring device 20, via the communication systems 66, 36, the start and end times of the immobility phases thus detected. Steps 1004 and 1008 are followed by a step 1100 of selecting periods of interest for the acquired parameters. During this step 1100, the processing unit 34 of the measuring device 20 identifies the periods between the start and end times of the immobility phases of the lower limbs as forming periods of interest for the acquired parameters and selects these periods. The accelerations ^(t. k) and rotation speeds ^^ (tk) acquired outside these periods are then discarded. Then, for each period of interest, the method 1000 comprises a step 1200 of determining a linear speed of the measurement point at various times during the period of interest. During this step 1200, the processing unit 34 determines the linear speed of the measurement point during certain phases. Said phases are, for example, phases of immobility of the measuring device 20 that the processing unit 34 detects and for which it determines a zero linear speed of the measurement point during these phases. These phases of immobility are typically detected by comparing the norm of the rotation speed ^^ (t k ) with a threshold, an immobility phase being detected when, for several sampling instants t ksuccessive, the standard of the rotation speed ^^ (tk) is lower than said threshold. Preferably, said phases also include phases of immobility of the elbow 28, which the processing unit 34 detects when the movement of the measuring point is a pure rotation of radius equal to the distance between the measuring point and the elbow 28. The linear speed of the measuring point is then easily determined, in a manner known to those skilled in the art. Advantageously, said phases also include phases during which the linear speed of the measuring point is determined by means of data other than that provided by the inertial device 33. These other data are for example made up of magnetic field gradient measurement data provided by the magnetometer network.The phases during which the linear speed of the measurement point is determined then include phases of stationarity of the ambient magnetic field, the processing unit 34 then implementing the method described in WO 2019 / 016474 to determine the linear speed of the measurement point during these phases. The phases during which the linear speed of the measurement point is determined are spaced from each other by time intervals during which the linear speed of the measurement point is not a priori known. The method 1000 comprises, for each of these intervals, a step 1300 of estimating an effort exerted by the muscles of the upper limb during this interval. This step 1300 is implemented by the processing unit 34. With reference to Figure 5, step 1300 comprises a first sub-step 1310 of estimating the evolution of an orientation of the measurement point during the interval.By "orientation of the measuring point" we mean here and in the following an orientation of the moving reference point (R. m ) in a terrestrial inertial frame (R i ) having an axis coincident with the vertical, such that the angular speed ^^ (tk) is equal to the rotation speed of the moving frame (R m ) relative to said inertial reference frame (R i ). Here and in the following, we define the inertial frame (R i) as a direct orthogonal reference frame formed by a triplet of axes, represented in Figure 1, comprising: - a Z axis coincident with the vertical, - a horizontal X axis, and - a horizontal Y axis forming with the X and Z axes a direct X, Y, Z orthogonal reference frame. The orientation of the measurement point can be given by a rotation matrix (noted R), an orientation quaternion (noted Q), or Euler angles (roll φ, pitch θ, yaw ψ). These three notations being equivalent, they are used interchangeably in this document. We know in particular that the formula giving the transition matrix from the inertial reference frame (Ri) to the mobile reference frame (Rm) from the Euler angles is written: With reference to Figure 6, the sub-step 1310 comprises a first sub-step 1312 of determining an initial orientation of the measurement point at an initial time t0 at the start of the interval. During this sub-step 1312, the processing unit 34 first determines a starting orientation of the measurement point at a starting time t0 at which the forearm 22 and therefore the measuring device 20 is considered immobile. This determination is typically carried out from the acceleration measurements, the measured acceleration ^(t0) being equal to the opposite of the gravitational field: ^(t0) = -^. The rotation matrix at the starting time t0 is then such that it solves the following equation: in which: - ^ is the vertical unit vector of the inertial frame (Ri), - ^ ( ^ ^ ) is the acceleration measured by the accelerometer 42 at the starting time t0, expressed in the moving frame (R m), and - ‖^ norm of the measured acceleration ^(^ ^ ) at time t0. Note that this equation is not completely deterministic and leaves degrees of freedom on certain components of the rotation matrix R(t0). This is explained by the fact that the yaw angle ψ(t0) at the starting time t0 is of little importance for the rest of the calculations and that the X axis can therefore be fixed arbitrarily. For example, the X axis is chosen so that the yaw angle ψ(t0) at the starting time t0 is zero, that is to say so that the rotation matrix at the starting time t0 solves the following additional equation: in which: - ^ ^ is the unit vector of the inertial frame (Ri) corresponding to the Y axis, - ^ ^ is the unit vector of the moving frame (Rm) corresponding to the U axis, - ^ ( ^ ^ ) is the acceleration measured by the accelerometer 42 at the starting time t0, expressed in the moving frame (Rm ), - ‖ ^ norm of the measured acceleration ^ ( ^ ^ ) at time t0, and - ∧ is the vector product operator. Preferably, the starting time t0 is equal to the initial time t i ; the initial orientation is then equal to the starting orientation. Alternatively, the starting time t0 is prior to the initial time ti, the initial orientation then being deduced from the starting orientation by forward integration of the measurements provided by the gyrometer 40. By “forward integration” is meant here and in the following that the integration is carried out in positive time, with an initial value chosen in the past. The sub-step 1312 is followed by a sub-step 1313 of forward integration of the measured orientation on the basis of the initial orientation. During this sub-step 1313, the processing unit 34 produces a first estimate R f of the orientation of the measuring point at each sampling time t kof the interval by integrating the rotation speed measurements prior to time tk, taking as initial value the initial orientation at time ti. Preferably, this determination comprises the implementation of a linear state estimator filter (Luenberger filter, Kalman filter, etc.) or non-linear (extended Kalman filter, invariant observer, etc.). In the present description, the implementation of an extended Kalman filter is described, but those skilled in the art will be able to transpose to other filters. The implementation of said extended Kalman filter firstly comprises a prediction step, during which the processing unit 34 uses the following formula: R ^ ( ^^ ) = R^ ( ^^^^ ) + Ṙ^ ( ^^ ) dt in which t k-1 is the sampling time immediately preceding the sampling time t k , is the time derivative of the rotation matrix of the moving frame (Rm) in the inertial frame (Ri), given by the following differential equation: 0 −ω^ ( ^^ ) ω^ ( ^^ ) Ṙ ^ (^ ^ ) = ^ ω^ ( ^^ ) 0 −ω^ ( ^^ ) ^ × R ^ (^ ^^^ ) −ω ^ (^ ^ ) ω ^ (^ ^ ) 0 Noting R ^ ^ the estimation of the matrix R ^ after n sample steps, the estimation of the matrix R ^ ^ for each measure East : The equation is given here for a rotation matrix but an equivalent differential equation exists between attitude quaternion and angular velocity. In parallel, the processing unit 34 modifies, by linearization of the differential equation, a covariance matrix estimating the covariance between each state of the filter. The prediction step is followed by an update step. This update step uses the fact that the acceleration γ^ measured by the accelerometer 42 is equal to the sum of the gravity field g^ and the acceleration ^ linked to the movements of the measuring device 20: γ^ = ^ + g^ We also know that the acceleration ^ is on average zero (we understand that the upper limb 12 regularly returns to its starting position and that the movements in one direction or the other statistically compensate each other). We deduce from this that the measured acceleration γ^ expressed in the terrestrial inertial frame (Ri) is equal on average to −g^ . From the rotation matrix R^ ^ and accelerometer measurements in the moving frame (Rm), we can express the measured acceleration γ^ in the inertial frame (Ri): The difference between the estimated gravity field and the actual Earth gravity field is then written: This difference, which must be zero on average, is used for the orientation recalibration during the update step. Furthermore, by linearizing this formula, we can calculate the Kalman gain and consequently update the state and the covariance matrix. The update of the covariance matrix is ​​based on the assumption that the errors due to the sensors and the approximations are modeled as Gaussian distribution noise. The variance is estimated by measuring the noise of the sensors at rest and from assumptions on the movements made by the forearm 22. We thus obtain a first estimate of the orientation of the measurement point at each sampling instant tk of the interval. Sub-step 1313 is followed by a sub-step 1314 of determining a final orientation of the measurement point at a final instant tf at the end of the interval.During this sub-step 1314, the processing unit 34 first determines a reference orientation of the measuring point at a reference time t. r during which the forearm 22 and therefore the measuring device 20 is considered immobile. This determination is typically carried out from the acceleration measurements, the measured acceleration γ(tr) being equal to the opposite of the gravitational field: ^(tr) = -^. The rotation matrix at the reference instant t r is then such that it solves the following equation: ^ ^ ( ^ ^ ( ^ ) ^ = ^ ) ‖ ^ ( ^^ )‖ in which: - ^ is the vertical unit vector of the inertial frame (R i ), - ^ ( ^ ^ ) is the acceleration measured by the accelerometer 42 at the reference instant tr, expressed in the moving frame (Rm), and - ‖ ^ ( ^ ^ )‖is the norm of the measured acceleration ^ ( ^ ^ ) at time t r This equation allows us to obtain the roll and pitch at the reference instant t r , which are given by the following formulas: in which γ ^ is the component of the measured acceleration ^ along the U axis, γ ^ is the component of the measured acceleration ^ along the V and γ axis ^ is the component of the measured acceleration ^ along the W axis. The yaw angle ψ(tr) is fixed so as to be equal to the yaw angle obtained by downstream integration of the rotation speed measurements from the starting time t0 to the reference time tr. Preferably, the reference time t r is equal to the final time t f ; the final orientation is then equal to the reference orientation. Alternatively, the reference time t r is later than the final time t f, the final orientation then being deduced from the reference orientation by backward integration of the measurements provided by the gyrometer 40. By "backward integration" is meant here and hereinafter that the integration is carried out in negative time, with an initial value chosen in the future. Sub-step 1314 is followed by a sub-step 1315 of backward integration of the measured orientation on the basis of the final orientation. During this sub-step 1315, the processing unit 34 produces a second estimate Rb of the orientation of the measurement point at each sampling instant tk of the interval by integration of the rotation speed measurements subsequent to instant tk, taking as initial value the final orientation at instant tf. Preferably, this determination includes the implementation of a linear state estimator filter (Luenberger filter, Kalman filter, etc.) or non-linear (extended Kalman filter, invariant observer, etc.).In the present description, the implementation of an extended Kalman filter is described, but those skilled in the art will be able to transpose it to other filters. This implementation is substantially identical to that of the downstream integration step 1313 described above, with the following differences: - the formula used in the prediction step is as follows: R. ^ ( ^^ ) = R^ ( ^^^^ ) − Ṙ^ ( ^^ ) dt in which t k+1 is the sampling instant immediately following the sampling instant tk, and - the differential equation giving the matrix Ṙ ^ ( ^ ^ ) instantaneous rotation of the moving frame (Rm) in the inertial frame (Ri) is expressed as follows: Ṙ ^ (^ ^ ) × R ^ (^ ^^^ ) In essence, these differences can be summarized as follows: whereas forward integration uses prior estimates for the prediction step of the extended Kalman filter, backward integration uses posterior estimates. This gives a second estimate of the orientation of the measurement point at each sampling time t k of the interval. Sub-step 1310 concludes with a sub-step 1316 of merging the upstream and downstream integrations, during which a final estimate of the orientation of the measurement point at each sampling instant tk of the interval is calculated by linear-spherical interpolation of the first and second estimates: where: - ^ ^ ( ^ ^ ) is the first estimate of the orientation obtained during sub-step 1313 by downstream integration, - ^ ^ (^ ^) is the second estimate of the orientation obtained during sub-step 1315 by backward integration, and - ^ is given by the following formula: ^ = ^ ^ ^^ ^ Returning to Figure 5, step 1300 comprises, in parallel with step 1310, a step 1320 of estimating a linear speed of the measurement point during the interval. With reference to Figure 7, this sub-step 1320 comprises a first sub-step 1322 of downstream integration of the displacement parameters on the basis of the linear speed at the start of the interval. During this sub-step 1322, the processing unit 34 produces a first estimate Vf of the linear speed of the measurement point at each sampling instant tk of the interval by integrating the acceleration and rotation speed measurements prior to instant t k , taking as initial value an initial speed at time t i. This initial speed is equal to the linear speed of the measurement point at the end of a phase for which the linear speed was determined during step 1200 and which immediately precedes the interval. The first estimate of the linear speed of the measurement point is estimated iteratively at each sampling time tk, starting from the oldest, using the following formula: in which ^ ^ ^^ ^ ( ^ ^ ) is the first estimate of the linear velocity at sampling time t k , is the first estimate of the linear velocity at a sampling time t k-1 immediately preceding the sampling time t k (and therefore estimated at the previous iteration), is the acceleration at sampling time t k , given by the following formula: ^ ^ ^^ ^ ̇ ( ^^ )= ^ ( ^^ ) where: - ^ ( ^ ^ ) is the acceleration measured by accelerometer 42 at time tk, expressed in the moving frame (Rm), - ^ ^^→^^ (^ ^ ) is the rotation matrix giving the orientation of the measurement point in the inertial frame, said rotation matrix being either equal to the first estimate Rb(tk) of the orientation at the sampling instant tk estimated during sub-step 1313, or equal to the final estimate R(tk) of the orientation at the sampling instant t k estimated during sub-step 1316, - ^ is the gravitational field expressed in the inertial frame (Ri), - ^^ (^ ^^^ ) is the rotation speed measured by the gyrometer 40 at time t k-1, expressed in the moving frame (Rm), and - ∧ is the vector product operator. This gives a first estimate of the linear speed of the measurement point during the interval. Sub-step 1322 is followed by a sub-step 1324 of backward integration of the displacement parameters on the basis of the linear speed at the end of the interval. During this sub-step 1324, the processing unit 34 produces a second estimate Vb of the linear speed of the measurement point at each sampling instant tk of the interval by integrating the acceleration and rotation speed measurements subsequent to instant t k , taking as initial value a final speed at time t f. This final speed is equal to the linear speed of the measurement point at the beginning of a phase for which the linear speed was determined during step 1200 and which immediately follows the interval. The second estimate of the linear speed of the measurement point is estimated iteratively at each sampling time tk, starting from the most recent, using the following formula: in which is the second estimate of the linear velocity at sampling time t k , is the second estimate of the linear velocity at a sampling time t k+1 immediately following the sampling time t k (and therefore estimated at the previous iteration), and is the acceleration at the sampling time tk, given by the following formula: where: - ^(^ ^ ) is the acceleration measured by accelerometer 42 at time t k, expressed in the moving frame (R m ), - ^ ^^→^^ ( ^ ^ ) is the rotation matrix giving the orientation of the measurement point in the inertial frame, said rotation matrix being either equal to the second estimate Rb(tk) of the orientation at the sampling instant tk estimated during sub-step 1315, or equal to the final estimate R(tk) of the orientation at the sampling instant tk estimated during sub-step 1316, - ^ is the gravitational field expressed in the inertial frame (R i ), and - ^^ ( ^ ^^^ )is the rotation speed measured by the gyrometer 40 at time tk+1, expressed in the moving frame (Rm), and - ∧ is the vector product operator. This gives a second estimate of the linear speed of the measurement point during the interval. Sub-step 1320 concludes with a sub-step 1326 of merging the upstream and downstream integrations, during which a final estimate of the linear speed of the measurement point at each sampling time t k of the interval is calculated by linear interpolation of the first and second estimates: where: - is the first estimate of the linear speed obtained during sub-step 1322 by downstream integration, - ^ ^^^^ ^ (^ ^ ) is the second estimate of the linear speed obtained during sub-step 1324 by backward integration, and - ^ is given by the following formula: ^ = ^ ^ ^^ ^ ^ ^ ^^ ^. We thus obtain an estimate of the evolution of the orientation and an estimate of the evolution of the linear speed of the measurement point during the interval. Returning to Figure 5, sub-steps 1310 and 1320 are followed by a sub-step 1330 of defining an inertial calculation frame of reference. This inertial calculation frame of reference is for example the same as the inertial frame of reference (R i ) used for estimating orientation and linear speed. Alternatively, this inertial reference frame for calculation is any other inertial reference frame and preferably a reference frame having: - a first Z axis coincident with the vertical, - a second X axis, horizontal, oriented in the direction in which the axis of the moving reference frame (R) was pointing m) collinear with the forearm 22 when said axis presented the most horizontal orientation of the interval, and - a third axis Y, horizontal, forming with the first and second axes a direct orthogonal reference X, Y, Z. The orientations R(tk) and the speeds ^ ^ ( ^ ^ )previously estimated are then recalculated to be expressed in said inertial calculation frame of reference, by means of a simple change of reference frame well known to those skilled in the art. The sub-step 1330 is itself followed by a sub-step 1340 of calculating candidate trajectories for the center of gravity of the forearm 22. This calculation is based on a biomechanical model 100 of the upper limb 12, represented in Figure 9. As visible in this Figure, the biomechanical model 100 comprises a first segment 102 having a free end 104. It also comprises a second segment 106 articulated by a first 108 of its ends 108, 110, via a first ball joint 112, to a proximal end 114 of the first segment 102 opposite the free end 104, and articulated by its second end 110, via a second ball joint 1116, to a fixed point. The first segment 102 models the forearm 22 of the upper limb 12.It is constituted by a cylinder having a radius r and a length l equal to the distance between the measuring device 20 and the elbow 28. It forms with the vertical Z a first primary angle α modeling an angle of the forearm 22 with the vertical, and with the plane (X, Z) a first secondary angle ζ (not shown) modeling an angle of the forearm 22 with said plane. The second segment 106 models the arm 25 of the upper limb 12. It is constituted by a cylinder having a length L equal to the length of said arm 25. It forms with the vertical a second primary angle β modeling an angle of the arm 25 with the vertical, and with the plane (X, Z) a second secondary angle η (not shown) modeling an angle of the arm 25 with said plane. Each of the segments 102, 106 is rigid, that is to say that it is not curvable. Preferably, at least one of the segments 102, 106 is extensible.If the two segments 102, 106 are extensible, these extensions are advantageously constrained so as to be equal to each other. The center of gravity CG of the forearm 22 is placed on the first segment 102, at a distance from the joint 112 substantially equal to half the length of the forearm 22. The biomechanical model 100 thus has a plurality of possible configurations, each defined by the values ​​of the first primary α and secondary ζ angles and the values ​​of the second primary β and secondary η angles. During sub-step 1340, a candidate trajectory is calculated for each of a plurality of pairs of starting and finishing configurations 130, 132 of the biomechanical model 100, shown in Figure 10, respectively modeling the configuration of the upper limb 12 at the start and end of the interval.By hypothesis, the first secondary angle ζ is defined in each of these configurations as being equal to the heading ψ of the measurement point, and the second secondary angle η is set equal to the first secondary angle ζ. Each of the starting and arrival configurations of the biomechanical model 100 is therefore directly defined by the first and second primary angles α, β. With reference to Figure 8, sub-step 1340 thus begins, for the purposes of defining the starting and arrival configurations, with a sub-step 1342 of deducing a starting value α. i (Figure 10) of the first primary angle α and a sub-step 1344 of deducing a value on arrival α f (Figure 10) of the first primary angle α. In sub-step 1342, the first primary angle at the start αi is deduced from the orientation of the measurement point at the initial time ti, using the following formula: ^ ^ = in which ^ is the vertical unit vector of the inertial frame (Ri) est le unit vector defining the direction of the forearm 22 in the inertial frame (Ri) at the initial instant ti, said vector given by the following formula: where R t (ti) is the transpose of the rotation matrix at the initial time ti and ^ ^ is the unit vector of the moving reference frame (Rm) corresponding to the U axis. During sub-step 1344, the first primary angle on arrival αf is deduced from the orientation of the measurement point at the final instant tf, using the following formula: ^ ^ = arccos^ ^ ^ ^^^ ^ . ^^ in which ^ is the unit vector defined above and ^ ^ ^^^ ^ is the unit vector defining the direction of the forearm 22 in the inertial frame (R i ) at the final time t f , said vector ^ ^ ^^^ ^given by the following formula: where R t (t f ) is the transpose of the rotation matrix at the final time t f and ^ ^ is the unit vector of the moving frame (R m ) corresponding to the axis U. These sub-steps 1342, 1344 are represented as being implemented one after the other. Alternatively, they are implemented in parallel to each other. The sub-steps 1342, 1344 are followed by a sub-step 1346 of calculating a displacement of the elbow 28. Alternatively, they are implemented in parallel with this sub-step 1346. During the sub-step 1346, the processing unit 34 calculates a displacement ∆ ^ of the elbow 28 between the initial instant ti and the final instant tf, from the speeds and orientations estimated during the sub-steps 1310, 1320. This displacement ∆ ^ is given by the following formula: Or ^ ^ ^^^(^^ ^ ^^)is the linear velocity vector estimated at time t et ^ ^ ^^^ ^ are the unit vectors defined above. Sub-step 1340 further comprises, following sub-step 1342, a sub-step 1348 of determining a range of starting values ​​of the second primary angle β compatible with the starting value αi of the first primary angle α. Preferably, this sub-step 1348 comprises: - the comparison of the norm of the displacement ∆ ^ with the length L of the arm 25, and - the comparison with the value 1 of the difference between the cosine of the initial value α i of the first primary angle α and the ratio of the displacement norm ∆ ^ on the length L of the arm 25 (cos We then have different scenarios: - if the displacement norm ∆ ^is less than or equal to twice the length L and the difference between the cosine of the starting value αi of the first primary angle α and the ratio of the displacement norm ∆ ^ on the length L of the arm 25 is less than or equal to 1, then the range of compatible starting values ​​is determined to be equal to the range [0, α i ] ; - if the displacement norm ∆ ^ is strictly greater than twice the length L or the difference between the cosine of the starting value αi of the first primary angle α and the ratio of the displacement norm ∆ ^ on the length L of the arm 25 is strictly greater than 1, then the range of compatible starting values ​​is determined to be equal to the range [0, α i+δ], where δ is a small angle of predetermined value, typically between 20° and 40° and for example substantially equal to 30°. Alternatively, substep 1348 comprises only one of these comparisons, the choice of the range [0, αi] or [0, αi+δ] then depending on the result of this single comparison. Alternatively, substep 1348 does not include any comparison and the range of compatible starting values ​​is directly determined as being equal to the range [0, α i ]. Sub-step 1348 is followed by a sub-step 1350 of constructing a starting configuration of the biomechanical model 100. This sub-step 1350 comprises the selection 1352 of a starting value βi of the second primary angle β within the range of compatible starting values. To construct the starting configuration of the biomechanical model 100, the first primary angle α is thus assigned the starting value α i and at the second primary angle β the starting value βi . Sub-step 1350 is followed by a sub-step 1354 of downstream integration of the estimated linear speed V on the basis of the initial configuration. During this sub-step 1354, the processing unit 34 produces a first estimate P f (t k ) of the position of the measuring point at each sampling time t k of the interval by integrating the estimates of the linear velocity V of the measurement point prior to time tk, taking as initial value an initial position Pi of the measurement point at time ti. This initial position is given by the following formula: in which: - L is the length of the arm 25, - l is the distance between the measuring device 20 and the elbow 28, - is the unit vector defining the direction of the forearm 22 in the inertial frame (R i ) at the initial time t i, - Qi is a rotation matrix corresponding to a rotation around the vertical axis Z converting the unit vector ^ into the horizontal projection of the vector The first estimate of the position of the measuring point is estimated at each sampling time t k using the following formula: ^ ^ (^ ^ ) A first estimate of the trajectory of the measurement point during the interval is thus obtained. In parallel with this sub-step 1354, the sub-step 1340 comprises a sub-step 1355 for determining a range of arrival values ​​of the second primary angle β compatible with the starting value βi of said angle β and / or with the arrival value αf of the first primary angle α. Preferably, this sub-step 1355 comprises: - the comparison of the norm of the displacement ∆ ^with the length L of the arm 25, and - the comparison with the value 1 of the sum between the cosine of the arrival value αf of the first primary angle α and the ratio of the displacement norm ∆ ^ on the length L of the arm 25 We then have different scenarios: - if the displacement norm ∆ ^ is less than or equal to twice the length L and the sum between the cosine of the arrival value αf of the first primary angle α and the ratio of the displacement norm ∆ ^ on the length L of the arm 25 is less than or equal to 1, then the range of compatible arrival values ​​is determined as being equal to an angular interval extending over less than 20°, preferably over less than 15°, for example over substantially 10°, centered on an angle βt verifying the following relationship: ^ ( cos ( ^ ^ ) − cos ( ^ ^ )) = ∆ ^ . ^ - if the displacement norm ∆^ is strictly greater than twice the length L or the sum between the cosine of the arrival value α f of the first primary angle α and the ratio of the displacement norm ∆ ^ on the length L of the arm 25 is strictly greater than 1, then the range of compatible arrival values ​​is determined to be equal to the range [0, α f +ε], where ε is a small angle of predetermined value, typically between 0° and 40°, preferably between 20° and 40° and for example substantially equal to 30°. Alternatively, substep 1355 comprises only one of these comparisons, the choice of the angular interval centered on the angle βt or of the range [0, αi+ε] then depending on the result of this single comparison. Alternatively, substep 1355 does not include any comparison and the range of compatible starting values ​​is directly determined as being equal to the angular interval centered on the angle β t. Sub-step 1355 is followed by a sub-step 1356 of constructing an arrival configuration of the biomechanical model 100. This sub-step 1356 comprises the selection 1357 of an arrival value βf of the second primary angle β within the range of compatible arrival values. To construct the arrival configuration of the biomechanical model 100, the arrival value α is thus assigned to the first primary angle α. f and at the second primary angle β the arrival value β f . Sub-step 1356 is followed by a sub-step 1358 of upstream integration of the estimated linear velocity V on the basis of the arrival configuration. During this sub-step 1358, the processing unit 34 produces a second estimate Pb(tk) of the position of the measurement point at each sampling instant tk of the interval by integrating the estimates of the linear velocity V of the measurement point subsequent to the instant t k, taking as initial value a final position P f from the measuring point to the final time t f . This initial position is given by the following formula: in which: - L is the length of the arm 25, - l is the distance between the measuring device 20 and the elbow 28, - ^ ^ ^^^ ^ is the unit vector defining the direction of the forearm 22 in the inertial frame (Ri) at the final instant tf, - Qf is a rotation matrix corresponding to a rotation around the vertical axis Z converting the unit vector ^ into the horizontal projection of the vector ^ ^ ^^^ ^ . The second estimate of the position of the measuring point is estimated at each sampling time t k using the following formula: ^ ^ ^ ^ (^ ^ ) = ^ ^ + ^ ^ ^ ^^^ ( ^^ ^ ^^ ) ^^ ^ ^A second estimate of the trajectory of the measurement point during the interval is thus obtained. Sub-step 1340 concludes with a sub-step 1359 of merging the upstream and downstream integrations, during which a final estimate of the position of the measurement point at each sampling instant tk of the interval is calculated by linear interpolation of the first and second estimates: ^(^ ^ ) where: - ^ ^ (^ ^ ) is the first estimate of the position obtained during sub-step 1354 by forward integration, - ^ ^ ( ^ ^ ) is the second estimate of the position obtained during sub-step 1356 by backward integration, and - ^ is given by the following formula: ^ = ^ ^ ^^ ^. A final estimate of the trajectory of the measurement point during the interval is thus obtained for the pair of departure and arrival configurations. From this final estimate of the trajectory of the measurement point, the processing unit 34 easily deduces a candidate trajectory of the center of gravity CG, this trajectory being composed of a set of candidate positions at time tk deduced from the final estimates of the positions of the measurement point by means of the following formula: ^ ^ ( ^^ ) in which d is the distance from the center of gravity C G at the measuring point and ^ ^^^^(^^^^^^ ^ ^^) is a unit vector defining the direction of the forearm 22 in the inertial frame (R i ) at time tk, said vector ^ ^ ^^^(^^^ ^ ^^ ^ ^^) given by the following formula: where R t (t k) is the transpose of the rotation matrix at sampling time t k and ^ ^ is the unit vector of the moving reference frame (Rm) corresponding to the U axis. Sub-steps 1356, 1358 and 1359 are repeated for each arrival value βf included in the range of compatible values ​​determined during sub-step 1355 and sub-steps 1350, 1354, 1355, 1356, 1358 and 1359 are repeated for each departure value β iincluded in the range of compatible values ​​determined during substep 1348. It will be noted that, by “each value included in the range”, we mean a set of discrete values ​​spaced from each other by a predetermined step, for example substantially equal to 0.1°. Thus, for each starting value βi included in the range of compatible values ​​determined during substep 1348 and for each arrival value βf included in the range of compatible values ​​determined during substep 1355, there is a pair of starting and arrival configurations for which a candidate trajectory of the center of gravity CG is calculated. Returning to Figure 5, sub-step 1340 is followed by a sub-step 1360 of evaluating a compatibility of each candidate trajectory with the biomechanical model 100. During this sub-step 1360, the processing unit 34 evaluates a compatibility of each candidate trajectory with the biomechanical model 100.This compatibility is typically evaluated by means of a cost function which is a function of the deformations that the tracking of the candidate trajectory imposes on the biomechanical model 100. Typically, the compatibility is inversely proportional to the result of the cost function: for example, in the case where the segment 106 (which corresponds to the arm 25) is extensible, the cost function can be the difference between the maximum extension of the segment 106 and the length L, the compatibility being inversely proportional to said difference. Alternatively, the cost function is constructed in such a way that the compatibility increases with the result of the cost function. The sub-step 1360 is followed by a sub-step 1365 of selecting the candidate trajectory having the best compatibility.During this sub-step 1365, the processing unit 34 selects the candidate trajectory whose result of the cost function reflects the best compatibility, that is to say whose result is the lowest (when the compatibility is inversely proportional to the result of the cost function) or the highest (when the compatibility increases with the result of the cost function). The sub-step 1365 is followed by a sub-step 1370 of verifying the acceptable nature of the compatibility of the selected trajectory. During this sub-step 1370, the processing unit 34 compares the result of the cost function of the selected trajectory with a predetermined threshold, said threshold being chosen in such a way that the comparison of the result of the cost function of the selected trajectory with said threshold reflects the acceptable nature or not of the compatibility of the selected trajectory.Typically, said threshold is chosen so that the compatibility of the selected trajectory is considered acceptable if and only if the result of the cost function is lower than the threshold (when the compatibility is inversely proportional to the result of the cost function) or higher than the threshold (when the compatibility increases with the result of the cost function). For each interval, which we will call in the following “primary interval”, for which the compatibility of the selected trajectory is acceptable, the sub-step 1370 is followed by a sub-step 1372 of validation of the selected trajectory. The selected trajectory therefore constitutes a validated estimate of the trajectory of the center of gravity C. Gduring said primary interval. Substep 1372 is followed by a substep 1374 of deducing a rotational component of the kinetic energy of the forearm 22 during the primary interval. During this substep 1374, the processing unit 34 deduces from the rotational speed measurements during the primary interval a rotational component Ek,r of a mass kinetic energy acquired by the forearm 22 during the primary interval. This rotational component is typically obtained using the following formula: in which: - ^ ̿ is the mass inertia tensor of the forearm 22, i.e. independent of the mass of the forearm 22, expressed in the moving frame (Rm), - ^^ (^) is the rotation speed of the measurement point at an instant t, expressed in the moving frame (R m) and typically provided by the gyroscope 40, and ^^^^ - ^^ (^) is the rotational acceleration of the measurement point at a time t, expressed in the moving frame (R m ) and typically deduced from the measurements of the gyroscope 40 by calculating the difference between two consecutive speed measurements. Sub-step 1374 is followed by a sub-step 1376 of evaluating a primary force Wp exerted by the upper limb 12 during the primary interval. During this step, the processing unit 34 evaluates said primary force Wp from the trajectory and the evolution of the estimated linear speed, typically by means of the following formula: ^ ^ = ^ ^,^ + ^ ^,^ + ^ ^ in which: - E k,ris the rotational component of the mass kinetic energy acquired by the forearm 22 during the primary interval mentioned above, - Ek,t is a translational component of the mass kinetic energy acquired by the forearm 22 during the primary interval, given by the following formula: in which: o is the linear speed of displacement of the center of gravity CG at a time t, expressed in the inertial frame (Ri) and deduced from the estimated linear speed of the measurement point ^ ^ (^) at time t, ^ ^ ^^^^^ o ^ ^^ (^) is the acceleration of the center of gravity CG at an instant t, expressed in the inertial frame (Ri) and typically deduced from the linear speed of displacement of the center of gravity CG by calculating the difference between the values ​​of said linear speed at two consecutive instants. - E pis a mass potential energy acquired by the forearm 22 during the primary interval, given by the following formula: in which: og is the norm of the gravitational field ^ oz(t f ) is the vertical coordinate of the center of gravity C G at the end of the primary interval, and oz(ti) is the vertical coordinate of the center of gravity CG at the beginning of the primary interval. Substep 1376 is further followed by a substep 1378 of determining a relationship f between primary force W p and rotational component E k,rof the kinetic energy acquired by the forearm 22 during each primary interval. This relationship f is typically determined by regression, preferably a linear regression, on the set of torques (primary force Wp, rotational component of the kinetic energy Ek,r) obtained for the different primary intervals. For each interval, which we will call in the following “secondary interval”, for which the compatibility of the selected trajectory is not acceptable, sub-step 1370 is followed by a sub-step 1382 of rejection of the candidate trajectories. None of the candidate trajectories is then considered as a valid estimate of the trajectory of the center of gravity C Gduring said secondary interval. Substep 1382 is followed by a substep 1384 of deducing a rotational component of the kinetic energy of the forearm 22 during the secondary interval. During this substep 1384, the processing unit 34 deduces from the rotational speed measurements during the secondary interval a rotational component Ek,r of the kinetic energy acquired by the forearm 22 during the secondary interval. This rotational component is typically obtained by means of the formula already mentioned for step 1374. Substep 1384 is followed by a substep 1386 of evaluating a secondary force Ws exerted by the upper limb 12 during the secondary interval. During this step, the processing unit 34 evaluates said primary force Ws from the rotational component Ek,r of the kinetic energy acquired by the forearm 22 during the secondary interval and from the relationship f between primary force Wp and rotational component Ek,r of the kinetic energy acquired by the forearm 22 determined during sub-step 1378. Typically, the processing unit 34 evaluates the primary force W s using the following formula: ^ ^ = ^(^ ^) Step 1300 is repeated for each time interval of the period of interest during which the linear speed of the measurement point is not a priori known and which is between two phases for which the linear speed of the measurement point was determined during step 1200. Steps 1200 and 1300 are repeated for each period of interest included in a predetermined period. This predetermined period typically has a duration greater than one week and, advantageously, less than three months. Following this repetition of steps 1200, 1300, the method 1000 comprises a step 1400 of determining the value of a statistical quantity, calculated on the set formed by the primary Wp and / or secondary Ws forces evaluated for the various time intervals included in the predetermined period, representative of a state of fitness of the person 14 over said predetermined period.During this step 1400, the processing unit 34 performs a statistical calculation on the set formed by the primary forces W. p and / or secondary W s evaluated for the various time intervals included in the predetermined period. Typically, the processing unit 34 calculates an average and / or a percentile of these efforts, relative to all the efforts evaluated for the predetermined period. For example, the processing unit 34 calculates the 50 e percentile, which corresponds to the median, the 80 e percentile, the 95th e percentile, the 99th e percentile, or any other percentile value of the efforts evaluated for the predetermined period. Preferably, the processing unit 34 selects the statistical quantity from a predetermined percentile range, in particular a high percentile range, i.e. percentiles greater than 70 epercentile. Indeed, these high percentiles are representative of the maximum effort and maximum muscular power that the person 14 is capable of developing, because they reflect the fastest and / or longest movements of an individual, the speed and duration of the movements of the upper limbs being constrained by the state of muscular strength. These high percentiles are therefore particularly sensitive to the state of fitness of the person 14. Advantageously, the statistical quantity is constituted by the 99 epercentile of the efforts evaluated for the predetermined period. Optionally, step 1400 comprises the selection of at least one other statistical quantity representative of a state of fitness of the person 14 over the predetermined period. Steps 1002 to 1400 are repeated for several predetermined periods. The processing unit 34 thus produces a set of values ​​of the statistical quantity representative of the state of fitness of the person 14 over several predetermined periods. Following this repetition of steps 1002 to 1400, the method 1000 comprises a step 1500 of observing the evolution of the statistical quantity across the different predetermined periods. The state of fitness of the person 14 can thus be monitored. Advantageously, this monitoring of the state of fitness of the person 14 is used to determine the effectiveness of a treatment administered to the person 14.Thanks to the exemplary embodiment described above, it is thus possible to monitor over time the physical fitness of an individual who cannot use his lower limbs, with few constraints for the individual. This monitoring is made possible in particular in an uncontrolled environment and over long periods, which allows for faithful monitoring of the actual fitness of the individual. The exemplary embodiment also allows for a good estimation, in an uncontrolled environment and over a long period, of the physical efforts made by an upper limb of a person for various movements. It further allows for the improvement of the quality of the estimations, in particular of linear speed and position, deduced from the measurements of an inertial device, in particular when this inertial device is attached to a limb of an individual.It will be noted that, although the above description focuses on an exemplary embodiment in which the measuring device 20 is attached to an upper limb 12 of a person 14, the forces evaluated being those of this upper limb 12, the invention also extends to a method for evaluating the forces of a lower limb of a person, the measuring device 20 then being attached to the leg of this lower limb. The invention also extends to a method for evaluating the forces of a foreleg or hindleg of an animal. Those skilled in the art will easily be able to make the necessary adaptations.

Claims

CLAIMS 1. Method for estimating a trajectory, over a movement interval, of a point of interest belonging to a limb (12) of an individual (14), the method being implemented by a data processing unit (34) and comprising the following steps: - estimation (1320) of the evolution, during the movement interval, of a speed of a measurement point integral with the point of interest, - estimation (1310) of the evolution of an orientation of the measurement point during the movement interval, and - estimation of the trajectory of the point of interest during the movement interval from the estimated speed and orientation of the measurement point.

2. Method according to claim 1, in which the individual (14) is a human being and the limb (12) is an upper limb of the individual (14), the point of interest preferably being a point of a forearm (22) of said upper limb. 3.Method according to claim 1 or 2, wherein the estimation of the trajectory comprises the following steps: - calculation (1340) of several candidate trajectories by applying the estimated speed and orientation to different pairs of starting and arrival configurations (130, 132) of a biomechanical model (100) of the limb (12), - evaluation (1360) of a compatibility of each candidate trajectory with the biomechanical model, and - selection (1365) of the candidate trajectory having the best compatibility.

4. Method according to claim 3, wherein the point of interest is a point of a forearm (22) belonging to an upper limb of the individual (14), each starting configuration (130) of the biomechanical model (100) of the limb (12) being defined by a pair of starting angles formed of a first starting angle (α. i) between the forearm (22) and the vertical and a second starting angle (βi) between the arm (25) and the vertical and each arrival configuration (132) of the model biomechanics (100) of the limb (12) being defined by a pair of arrival angles formed by a first arrival angle (α f) between the forearm (22) and the vertical and a second arrival angle (βf) between the arm (25) and the vertical.

5. Method according to claim 4, wherein, for each starting (130) or arrival (132) configuration of the biomechanical model (100) of the limb (12): - the first starting angle (αi), respectively arrival angle (αf), is deduced from the orientation of the measuring point at the start, respectively at the end, of the displacement interval, and - the second starting angle (βi), respectively arrival angle (βf), is less than the first starting angle (αi), respectively arrival angle (αf).

6. Method according to claim 4 or 5, wherein there exists, for each value of second starting angle (βi) less than the first starting angle (αi) and each value of second arrival angle (β f) included in a predetermined interval, a pair of departure and arrival configurations (130, 132) for which a candidate trajectory is calculated.

7. Method according to claim 6, in which the predetermined interval is constituted by an interval between 0° and the value of the first arrival angle (αf), or by an angular interval around an angle βt verifying the following relation: ^ ( cos in which: - L is the length of the arm (25) of the upper limb, - β i is the value of the second starting angle (β i ) in the starting configuration (130) belonging to the same pair as the arrival configuration (132), - ti is the start time of the displacement interval, - t f is the end time of the displacement interval, - ^ ^ ^^^ ( ^^ ^ ^^) is the velocity vector of the measuring point at each instant t of the displacement interval, - l is the distance between the elbow (28) of the upper limb and the measuring point, - ^ ^ ^^^ ^ is a unit vector giving the orientation of a principal axis of the forearm (22) at the end of the displacement interval, - ^ ^ ^^ ^ is a unit vector giving the orientation of the main axis of the forearm (22) at the start of the displacement interval, - ^ is a unit vector giving the vertical orientation, the vectors ^ ^^^ ( ^^ ^ ^^ ) , ^ ^^^^ ^ , and ^ being expressed in the same inertial frame of reference.

8. Method according to any one of claims 3 to 7, in which the calculation (1340) of each candidate trajectory comprises the following steps: - downstream integration (1354) of the estimated speed, taking as initial position the position of the measurement point in the departure configuration (130), - upstream integration (1358) of the estimated speed, taking as initial position the position of the measurement point in the arrival configuration (132), and - calculation (1359) of the candidate trajectory by merging the upstream and downstream integrations. 9.Method according to any one of the preceding claims, in which the estimation of the evolution of the speed of the measurement point during the movement interval comprises the following steps: - determination (1200) of an initial speed and a final speed of the measurement point respectively at the start and at the end of the movement interval, - estimation (1320) of the evolution of a speed of the measurement point during the movement interval, said estimation comprising:. ^ the downstream integration (1322) of displacement parameters of the measurement point acquired by an inertial device (32) during the displacement interval, taking as the initial value of the speed the value of the initial speed, ^ the upstream integration (1324) of the displacement parameters, taking as the initial value of the speed the value of the final speed, and ^ the calculation (1326) of the estimated speed by merging the upstream and downstream integrations. 10.Method (1300) for evaluating an effort exerted by a limb (12) of an individual (14), the method being implemented by a data processing unit (34) and comprising the following steps: - estimation of a trajectory of a point of interest belonging to the limb (12) over at least one primary movement interval by means of a method according to any one of the preceding claims, - evaluation (1376), from the estimated trajectory and speed evolution, of a primary effort exerted by the limb (12) during the or each primary movement interval. 11.Method (1300) according to claim 10, comprising the following additional steps: - deduction (1374), from the displacement parameters acquired during the primary displacement interval, of a rotational component of a primary kinetic energy acquired by at least a portion of the member (12) during the primary displacement interval, the primary force being a function of said rotational component of the primary kinetic energy, - determination (1378) of a relationship between the primary force and the rotational component of the primary kinetic energy, - estimation (1310) of the evolution of an orientation of a measurement point integral with the point of interest during at least one secondary displacement interval,. - deduction, from displacement parameters acquired by the inertial device (32) during a secondary displacement interval, of a rotational component of a secondary kinetic energy acquired by said at least one portion of the member (12) during the secondary displacement interval, - evaluation, from the determined relationship and the rotational component of the secondary kinetic energy, of a secondary force exerted by the member (12) during the or each secondary displacement interval. 12.

13. Method (1300) according to claim 11 comprising, for each of the primary and secondary displacement intervals: - the estimation (1320) of the evolution, during the primary or secondary displacement interval, of a speed of a measurement point integral with the point of interest, - the calculation (1340) of several candidate trajectories by applying the estimated speed and orientation to different pairs of departure and arrival configurations (130, 132) of a biomechanical model (100) of the limb (12), - the evaluation (1360) of a compatibility of each candidate trajectory with the biomechanical model (100), and - the verification (1370), for each primary or secondary displacement interval, of the acceptable nature of the compatibility of at least one candidate trajectory, the trajectory of the point of interest being estimated only when at least one candidate trajectory has an acceptable compatibility.Method (1000) for analyzing an effort exerted by a limb (12) of an individual (14), the method being implemented by a data processing unit (34) and comprising the following steps: - a) estimation, for several intervals of movement of a point of interest of the limb (12) occurring during a predetermined period, of an effort. exerted by said member (12) by means of a method (1300) according to any one of claims 10 to 12, - b) determining (1400) the value of at least one statistical quantity, calculated on the set formed by the estimated efforts, representative of a state of fitness of the individual (14).

14. Method (1000) according to claim 13, comprising repeating steps a) and b) for several predetermined periods and observing (1500) the evolution of the or each statistical quantity through the different predetermined periods.

15. Method for monitoring the state of fitness of an individual, comprising analyzing an effort exerted by a member of the individual by means of an analysis method according to claim 13 or 14 and deducing a state of fitness of the individual from said analysis.

16. The method of claim 15, comprising administering a treatment to the individual, and evaluating an effectiveness of the treatment from the inferred fitness status. 17.Equipment (10) comprising an inertial device (32) for acquiring displacement parameters of a measurement point during at least one displacement interval, and a data processing unit (34) configured to estimate, from the displacement parameters acquired by the inertial device (32), by implementing a method according to any one of claims 1 to 12, at least one of: a trajectory, over the displacement interval, of a point of interest integral with the measurement point, and a force exerted by a limb (12) of an individual (14) to which the point of interest belongs.

18. Computer program product comprising code instructions for implementing a method according to any one of claims 1 to 16 when said program is executed by a computer.

Images