Method for determining an internal muscular load, device and corresponding program.

Abstract

The invention relates to a method for determining data representative of an internal muscular load ( ) developed by at least one muscle belonging to a muscle group. The method comprises the steps: obtaining (A10) kinematic data (i(t)) and electromyographic data (emg(t)) collected during a contraction of said at least one muscle; determining (A20) a musculotendinous length ( ; determining (A30) a muscle activation level (a(t)); determining (A40) a muscle length and a contraction speed; determining (A50) an estimate of a muscular deformation (dM(t)); calculating (A60) a muscular stress index); and calculating (A70) the internal muscular load by the product of the muscular stress index and a duration (t) of said effort. Figure for the abstract: Fig.1

Description

Title of the invention: Method for determining an internal muscular load, device and corresponding program.

[0001] Domain

[0002] The present invention relates to the field of physiological measurement, and more particularly to the non-invasive physiological measurement of muscular effort associated with the performance of a motor task. Physiological measurement covers measurements of physical quantities resulting from the effort provided by the muscles of a user, in particular an internal muscular load of one or more muscles of the user.

[0003] The invention finds its application in different fields to evaluate the relationship between the muscular effort provided by the individual and the motor task performed. The invention can find use in the clinical field, ergonomics, and sport. Prior Art

[0004] The rapid rise of connected wearable devices such as watches or heart rate monitor belts is no longer really a topic of debate. Most individuals use a connected watch or an equivalent device during their daily lives and during their physical activities. Connected watches measure the external training load (e.g., speed and distance traveled * duration of effort), and the internal cardiovascular load (what the training cost the person who performed it; for example, by measuring heart rate * duration of effort). However, it is well known that the internal cardiovascular load estimated using heart rate is not related to the internal muscular load. In other words, it is not possible to estimate the intensity of muscular effort, especially of different muscles / muscle groups, using heart rate measurement alone.

[0005] To date, internal muscle load has not been estimated in the laboratory or in relevant environments (sport, medicine, work environment and in particular industrial, activities of daily life). However, quantifying internal muscle load would be useful on the one hand, to identify changes in acute and / or chronic muscle capacities of an individual (e.g., aging, acute / chronic fatigue associated with repetitive motor tasks performed in a professional / sporting context) and on the other hand to quantify whether the rehabilitation content proposed by health professionals effectively uses the targeted muscles or muscle groups (e.g., physiotherapist, rehabilitation physician, etc.). These findings are all the more relevant since a series of studies demonstrate that each individual has a unique way (i.e., muscle coordination) of using their muscles during tasks. varied motor skills (walking, pedaling, strengthening exercise). Depending on the muscular coordination used by the individual, the distribution and amplitude of adaptations (eg, hypertrophy) / alterations (eg, MSD, sarcopenia) may therefore vary widely from one individual to another.

[0006] It is thus noted that there is a need to provide tools for evaluating internal muscular load, in order to better understand the problems posed by the prior art. Summary of the invention

[0007] The technique described was designed taking into account these problems of the prior art. More particularly, the invention relates to a method for determining data representative of an internal muscular load developed by at least one muscle belonging to a muscle group, a method implemented by means of an electronic device comprising a processor and a memory. Such a method comprises the steps of: - obtaining kinematic data and electromyographic data collected during a contraction of said at least one muscle; - determination, for said at least one muscle, from the kinematic data, of data representative of a musculotendinous length, using a musculoskeletal model; - determination, for said at least one muscle, from the electromyographic data, of data representative of a level of muscular activation; - determination, for said at least one muscle, from the musculotendinous length and the level of muscular activation and two neuromuscular and musculotendinous models, of a muscular length and data representative of a contraction speed; - determination, for said at least one muscle, of an estimate of a muscular deformation, from the muscular length; - calculation of a muscle stress index, from muscle activation data, muscle deformation and a muscle strength indicator dependent on contraction speed; and - calculation of the internal muscular load by the product of the muscular constraint index and the duration of said effort.

[0008] Thus, it is possible to quantify the internal muscular load of a muscle or a muscle group in the context of a motor task. This internal muscular load characterizes, for a given external load, the level of effort provided by the muscle.

[0009] By calculating the internal muscle load, through the use of data muscle activation, muscle deformation and a muscle strength indicator dependent on the contraction speed, the method according to the invention makes it possible to predict more precisely than in the prior art the muscular alterations and adaptations when the muscle contracts at lengths greater than the optimal length.

[0010] According to a particular characteristic, the step of determining the musculotendinous length of said at least one muscle comprises the sub-steps of: - simulation of joint angles of joints moved during exercise, based on kinematic data and the musculoskeletal geometric model; and - calculation, from joint angles and musculoskeletal geometry described by a musculoskeletal model, of the temporal evolution of musculotendinous length.

[0011] According to a particular characteristic, the step of determining the data representative of the level of muscular activation of said at least one muscle comprises the sub-steps of: - filtering, in the frequency domain, of electromyographic data delivering signal envelopes; - rectification of signal envelopes; - normalization of signal envelopes, delivering normalized envelopes describing the level of nervous excitation of the muscle; and - calculation, from the level of nervous excitation of the muscle, of the data representative of muscular activation.

[0012] According to another aspect, the invention also relates to an electronic device, comprising a processor and a memory arranged for the determination of data representative of an internal muscular load developed by at least one muscle belonging to a muscular group. Such a device comprises: - means for obtaining kinematic data and electromyographic data collected during a static or dynamic effort of said at least one muscle; - means for determining, for said at least one muscle, from the kinematic and electromyographic data, a musculotendinous length, data representative of a level of muscular activation, data representative of a muscular length, an estimation of a muscular deformation, and data representative of a contraction speed; - calculation means, configured to calculate a muscle stress index from muscle activation and deformation data muscle and a muscle strength indicator dependent on the contraction speed, and to calculate data representative of the internal muscle load.

[0013] According to another aspect, the invention also relates to an electronic system comprising at least one processor and at least one memory arranged for the determination of data representative of an internal muscular load developed by at least one muscle belonging to a muscle group, said system comprising a smart garment, paired with a communication device connected to a processing server through a communication network. Such a system further comprises: - means for obtaining kinematic data and electromyographic data collected during a static or dynamic effort of said at least one muscle in contact with the smart garment; - means for determining, for said at least one muscle, from the kinematic and electromyographic data, a musculotendinous length, data representative of an activation level, data representative of a muscle length, an estimation of a muscle deformation and data representative of a contraction speed; - calculation means, configured to calculate a muscle stress index from muscle activation data, muscle deformation and an indicator fFV(t) of force dependent on the contraction speed, and to calculate data representative of the internal muscle load.

[0014] The means of obtaining, the means of determining and the means of calculating are distributed between the different components of the system.

[0015] According to a preferred implementation, the different steps of the methods according to the present disclosure are implemented by one or more software or computer programs, comprising software instructions intended to be executed by a data processor of an execution terminal according to the present technique and being designed to control the execution of the different steps of the methods, implemented at the level of a communication terminal, a remote server and / or a blockchain, within the framework of a distribution of the processing operations to be carried out and determined by a scripted source code or a compiled code.

[0016] Consequently, the present technique also relates to programs, capable of being executed by a computer or by a data processor, these programs comprising instructions for controlling the execution of the steps of the methods as mentioned above.

[0017] A program may use any programming language, and be in the form of source code, object code, or code intermediate between source code and object code, such as in a partially compiled form, or in any other desirable form.

[0018] The present technique also aims at an information medium readable by a data processor, and comprising instructions of a program as mentioned above.

[0019] The information medium may be any entity or terminal capable of storing the program. For example, the medium may comprise a storage means, such as a ROM (Read-Only Memory), for example a CD ROM or a microelectronic circuit ROM, or a magnetic recording means, for example a mobile medium (memory card) or a hard disk or an SSD.

[0020] On the other hand, the information carrier may be a transmissible carrier such as an electrical or optical signal, which may be conveyed via an electrical or optical cable, by radio or by other means. The program according to the present technique may in particular be downloaded over an Internet-type network.

[0021] Alternatively, the information carrier may be an integrated circuit in which the program is incorporated, the circuit being adapted to execute or to be used in the execution of the method in question.

[0022] According to one embodiment, the present technique is implemented by means of software and / or hardware components. In this regard, the term "module" may correspond in this document to a software component, a hardware component or a set of hardware and software components.

[0023] A software component corresponds to one or more computer programs, one or more sub-programs of a computer program, or more generally to any element of a program or software capable of implementing a function or a set of functions, according to what is described below for the module concerned. Such a software component is executed by a data processor of a physical entity (terminal, server, gateway, set-top-box, router, etc.) and is capable of accessing the hardware resources of this physical entity (memories, recording media, communication buses, electronic input / output cards, user interfaces, etc.).

[0024] Similarly, a hardware component corresponds to any element of a hardware assembly capable of implementing a function or a set of functions, as described below for the module concerned. It may be a programmable hardware component or one with an integrated processor for executing software, for example an integrated circuit, a smart card, a memory card, an electronic card for executing firmware, etc.

[0025] Each component of the system described above of course implements its own software modules.

[0026] The different embodiments mentioned above can be combined with each other for the implementation of the present technique.

[0027] Other aims, characteristics and advantages of the invention will appear more clearly on reading the following description, given as a simple illustrative, and non-limiting, example, in relation to the figures, among which: - [Fig.l] represents a schematic flowchart of the steps of the method according to the present invention; - [Fig.2] represents a second schematic flowchart of the steps of the method according to the present invention; - [Fig.3] graphically represents an example of the temporal evolution of the 1MT quantities of the semimembranosus, semitendinosus and biceps femoris muscles in meters during a running cycle; - [Fig.4a] represents an example of the evolution of force-speed relationships; - [Fig.4b] represents an example of force length evolution; - [Fig.5] represents a simplified physical architecture of a device for determining internal muscular load; and - [Fig.6] represents a simplified physical architecture of an internal muscle load determination system.

[0028] Description of an embodiment

[0029] Reminders of the principle

[0030] As stated previously, an object of the invention consists in the determination, for a given exercise, kinematics or set of exercises, of data (index) representative of the internal muscular load produced during this exercise or this set of exercises. This internal muscular load is opposed to the external training load, which represents all the mechanical productions associated with the motor tasks carried out by an athlete (eg, a distance covered at a given speed, a load lifted, etc.). The internal muscular load therefore depends on the individual who carries out the proposed training activity, unlike the external training load which does not consider the individual. To know this load, it is therefore necessary to determine the muscular effort and therefore to have data making it possible to quantify this muscular effort.

[0031] The disclosure therefore relates to a methodology for obtaining the internal muscular load, and by rebound on the determination of the muscular effort provided, at the muscle scale. The inventors determined that it was possible to calculate the muscular effort from non-invasive “external” sensors (i.e. not requiring implantation) and that these sensors could also be integrated within a textile or clothing. Two types of data are used to obtain the internal muscle load: electromyographic data and inertial kinematic data (for example, inertial data from inertial measurement units).

[0032] The disclosure thus also relates to a textile equipped with electromyographic type electrodes adapted to estimate muscle activation, and inertial units to access joint angles and speeds. The disclosure also relates to a method for carrying out a combination of the data obtained via this textile with data, to calculate a composite data item called "muscular constraint" (or muscular effort). Finally, the method developed derives another composite indicator, called "muscular internal load", which is the product of the "muscular constraint" by the duration of the effort during the exercise or set of exercises (muscular effort x the time during which the effort occurs).

[0033] The method implemented, described in relation to [Fig.l], comprises the steps of: - obtaining A10 of kinematic data i(t) and electromyographic data emg(t) collected during a static or dynamic effort of at least one muscle. The data are obtained from sensors (for example from the clothing worn by the individual). The kinematic data i(t) and electromyographic data emg(t) are for example obtained by means of inertial units (for inertial data) and electromyographic electrodes; - determination A20, for said at least one muscle, from the kinematic data (i(t)), of data representative of a musculotendinous length lM1(t), using a musculoskeletal model; - determination A30, for said at least one muscle, from the electromyographic data emg(t), of data representative of a level of muscular activation a(t); - determination A40, for said at least one muscle, from the musculotendinous length 1^(t), the muscle activation level a(t) and two neuromuscular and musculotendinous models Ml, M2, a muscle length lM(t) and data representative of a contraction speed vM(t); - determination A50, for said at least one muscle, of an estimate of a muscle deformation dM(t), from the muscle length lM(t); - calculation A60 of a muscle stress index (S^t)), from muscle activation data a(t), muscle deformation dM(t) and an in muscle force indicator (fFV(t)) dependent on contraction speed vM(t); - calculation A70 of the internal muscle load CM by the product of the muscle stress index sM(t) and a duration t of said effort. This internal muscle load can be calculated for a single muscle M or for a muscle group, depending on the electromyographic data emg(t) and the number of muscles covered and / or treated by the surface electrodes providing the electromyographic data emg(t).

[0034] More particularly, in one example, the method consists in determining the internal training load CM( t) specific to the muscles participating in a muscle group analyzed from kinematic data i( t ) and electromyographic data emg(t ) collected during a static or dynamic effort. More particularly, the integration of these data captured from the garment is implemented in the following manner. The kinematic data i( t ), collected during the effort, are used for example with a musculoskeletal geometric model (MSQ) to estimate the general skeletal kinematics, and obtain the musculotendinous lengths lMT( t) of the muscle-tendon systems within the muscle group analyzed, as described for example in [Al Borno et al., 2022]. This kinematics is estimated because we know, a priori, the rigid segment to which each inertial unit is connected.For inertial data, the relative position of the coordinate axes of the inertial units can then be associated with those of the rigid musculoskeletal segments. In another implementation example, the kinematic data can be replaced by data from resistance goniometers, placed within the garment, and whose resistance, for a given constant input current, varies according to the angle of these goniometers.

[0035] The EMG data emg(t) are subjected to a digital transformation to estimate the level of nervous excitation e(t) received by each of the muscles within the muscle group analyzed.

[0036] Neuromuscular and musculotendinous models are used to estimate the time evolutions of activation levels a(t), relative muscle lengths of relative muscle contraction speeds yM(t}, and a muscle strength indicator Ffv(t) dependent only on the contraction speed yM(t) of the muscles within the muscle group analyzed from the previously obtained data e(t) and lMT(t}.

[0037] Approaches are proposed, within the framework of this patent application, to make this method generic and the quantities it predicts specific to the individual and the muscle analyzed.

[0038] Finally, the variables a(t), lM(t) and FFV(t) estimated for and specifically for each muscle within the analyzed muscle group are combined to obtain the internal muscle training load sM(t) undergone by each of these muscles.

[0039] [Fig.2] shows the different stages of the general method described previously.

[0040] The data i(jj and emg( t) obtained (from the IMU and EMG) for a muscle group (comprising for example two muscles) are processed (MSQ, Processing) independently to obtain the nerve excitation levels e 1 ( t), e2 ( / ) and the musculotendinous lengths l^IT ( t), 12MT( t ) of the muscles constituting this muscle group. Generic or specific neuromuscular and musculotendinous models (Ml, M2) transform these variables into estimates of the activation level a2(t), muscle length l2M(t), and speed of contraction v2M( / ) of the muscle, for each of the muscles analyzed (in [Fig.l], two muscles are represented). Finally, these three estimated variables (activation level, muscle length and contraction speed) are combined (MCM1, MCM2) to estimate the internal load of each of the muscles analyzed.

[0041] More particularly, for the estimation of the internal training load, an association is made of the activation variables a(t), of muscular deformation dM (t) and of an indicator of muscular force fFV(t) dependent on the contraction speed VM(t) making it possible to estimate the indicator of muscular constraints. The deformation muscle dM(t) is determined from the muscle length by the following calculation / 'w( t)-loF, where lOpt is the optimal muscle length. The calculation of The muscle strength indicator is described further.

[0042] Finally, the indicator of muscular constraints is multiplied by the duration of effort to obtain the internal muscular load, as (JM — y (%) • / ^min)-

[0043] The value of s(%) can be determined in several ways. It can be an index of muscular constraints s(t), over the duration of the exercise. It can be a percentage relative to the indicator of maximum muscular constraints (or muscular stress) (which is estimated for example relative to the maximum over the exercise). Other determinations of s(%) are possible depending on the circumstances, in particular the physical activity or the desired goal.

[0044] In this example, the technique integrates the kinematic data into a MSQ (simulation) model, and the EMG data into a mathematical model to infer the biomechanical constraints of the muscles (i.e., the relative length of the muscles during contraction and the relative velocities of contractions), combined with the relative activation of muscles and over time, in order to provide data representative of the internal muscle load. Estimating the internal muscle load by implementing the described method (activation signatures and kinematics in a numerical model) could provide a better understanding of the origin of inter-individual differences in the distribution of muscle damage and hypertrophy. Finally, monitoring the internal muscle load during each training period is an indirect way to estimate changes in muscle strength (e.g. a decrease in internal muscle load for a given external load).

[0045] In the following paragraphs are described different examples of carrying out the data processing and calculations implemented to arrive at the internal muscular load. It is understood that the person skilled in the art will be able to make adaptations of calculations to fit the conditions of the species, according to the chosen implementation. In particular, the use of a musculoskeletal model is not necessarily supported by the use of a simulation platform and can be directly integrated within a processing software module which performs simplified calculations, for example according to variability parameters, such as an evaluation of the length of the different bones of the user's skeleton, an evaluation which can be made by the user himself, for example during an installation, on a communication device, of a software application to carry out all or part of the processing for determining the internal muscular load.

[0046] Processing of kinematic data and determination of iMT musculotendinous lengths

[0047] In this example, a methodology for obtaining the musculoskeletal length for each of the muscles analyzed is presented. In this example, an estimate of lM1{ t) is obtained by combining kinematic data iÇt) (here inertial data) with musculoskeletal modeling (MSQ).

[0048] The raw inertial data ihria(t) are obtained with inertial measurement units (IMUs) containing, in a non-limiting manner, an accelerometer, a gyroscope, and a magnetometer, all three three-dimensional. The raw data ibnit(t) are combined, processed, and transformed by sensor fusion in order to obtain for each IMU the temporal evolution of its rotation matrices in the form of quaternions i( t). The sensor fusion algorithms are for example provided by the suppliers of certain IMUs (XSens™ or APDM™).

[0049] One inertial unit is desirable per segment analyzed. For example, if the internal load of a muscle group of the right thigh is quantified with the method in Figure 1, including muscles inserting on the pelvis and the tibia, three IMUs are per example placed within the garment (worn by the user) so that they are positioned on the pelvis, the right thigh, and the right calf, and three series of kinematic variable data ipe1vis ( t ), icuisse ( t ), and i(ibia ( t ) are obtained. These three variables (series of variables) are used to reproduce the skeletal pose of the user during the exercise time.

[0050] Thus, it is possible to virtually reproduce the actual skeletal kinematics of the individual, performed during the studied effort by carrying out an inverse kinematics analysis from the kinematic data / ( / ). To correlate the history of the orientations of the IMUs with angular kinematics a(f) of the skeletal system, a skeletal computer model representing (specifically or generically) the user's skeleton is used. This skeletal model models its joint dynamics during the effort provided. A platform can be used for this modeling, for example the Opensim platform (biomechanical modeling, simulation and analysis platform). In operational conditions, the calculations carried out under such a platform can be directly integrated into the calculation device.

[0051] First, a calibration step adjusts the orientation of the IMUs and virtually represents them in the orthogonal coordinate system of the associated virtual bone segment, removing the possible constant offset existing between the real IMU and the virtual segment. This step is implemented by a short static acquisition.

[0052] Once the power plants are calibrated, it is possible to perform inverse kinematics simulations from the inertial data / ( / ). By minimizing the difference at each time step between the orientation of the bone segments of the MSQ model and the experimental quatemions i( tj, the real angular kinematics of the subject are virtually reproduced with the MSQ model. For example, if the internal load of a muscle group of the right thigh is studied with the method in Figure 1, including muscles inserting on the pelvis and the tibia, the data from the three IMUs make it possible to determine the angular kinematics a(t) of the hip and knee joints with fidelity and without drift. Software solutions make it possible to perform these calculations. [Al Bomo et al., 2022] describes an example of such a software solution (OpenSense) making it possible to obtain the angular kinematics a(t) of the lower limbs.

[0053] In order to obtain the temporal evolution of the musculotendinous lengths lMT(t) of the muscles analyzed during the exercise, we use the relationships between angular skeletal kinematics a(j) and musculotendinous lengths, ie, use (materialize) relationships of type (a) ■ This is possible by describing (parameterizing) the muscular geometries within a skeletal model (generic or adapted to To do this, muscles can be defined as unidirectional actuators, attached to bone structures with two insertion and termination points, and, if necessary, additional crossing points.

[0054] The length of the bone segments, their respective pose provide the angular kinematics and the placement of the definition points of the musculotendinous systems on the bone segments define in three dimensions the musculotendinous lengths lM, {j ). For example, Figure 3 reports an example of temporal evolution of the quantities lMT of the semimembranous (light gray, dotted), semitendinous (black, dotted), and biceps femoris (dark gray, solid line) muscles in meters during a running cycle of seven strides.

[0055] Whatever the calculation method actually used, the temporal evolution of the musculotendinous lengths of muscles belonging to a muscle group studied is estimated, from kinematic data tj recorded by IMUs during a static or dynamic effort. The quantities lMT are estimated geometrically (for example via the MSQ model), whose joint kinematics are defined in inverse kinematics by the experimental kinematic data [( t), and a geometric unidirectional definition of the muscles studied.

[0056] Processing of electromyographic data and determination of the level of nervous excitation e( t )

[0057] As explained previously, the electromyographic (EMG) data emg(t) provides the level of nervous excitation e(f) of the muscles studied.

[0058] A greater or lesser number of EMG electrodes is placed on the muscle group studied; typically, they can be integrated into the clothing worn by the user. The surface area and the shape of the clothing (i.e., bra, shorts, leggings, t-shirt) determine the number of electrodes used. Depending on the operational conditions (for example, depending on the clothing or depending on the area studied), between 2 and 256 electrodes can be used and placed a few millimeters or a few centimeters apart from each other. According to a particular embodiment, the assembly of these electrodes can be bipolar. This type of assembly makes it possible to obtain a representative measurement of the muscular excitation of a muscle or a muscle group. These electrodes collect the superficial myoelectric activity emg(t) of the muscles covered by the electrodes during the effort provided.

[0059] In order to eliminate the components of the collected EMG signals that do not influence force generation, the raw EMG signals emg( t ) are filtered in the frequency domain. The resulting envelopes are rectified. Finally, the resulting signals are normalized by the maximum amplitude of the myoelectric signals recorded during maximum voluntary contractions. The emg(t) signals thus processed provide levels of nervous excitation ¢( / ) of the muscles covered by these electrodes.

[0060] Determination of activation levels a(t), relative muscle lengths and relative muscle contraction speeds

[0061] Once the temporal evolutions of the quantities of nervous excitation e ( t ) and musculotendinous lengths ]MT ( t ) are obtained from the EMG and IMUs data, they are integrated into neuromuscular and musculotendinous mathematical models in order to determine the three variables necessary for the estimation of the training load: a(t), lM( / ) and vM(t)•

[0062] The inventors have determined that it is possible to estimate these variables with a Hill muscle model. In this case, a nonlinear elastic passive element, representing the mechanical properties of the tendon, is modeled in mathematical equilibrium with an active nonlinear element, representing the contractile mechanics of the muscle.

[0063] The tendon is defined by its length [T and its force pT, and the active muscular element by its length lM, its contraction speed vM (both sought here), and its force PM. The following two relationships are used to identify the values ​​taken by the three sought variables:

[0064]

[0065] FT(t)=FM(t)

[0066] The force FT developed passively by the tendon follows an exponential force-deformation law, whose stiffness coefficient and the non-force transmission length f are specific to the muscle studied, and for which charts are available or experimental measurements are possible.

[0067] The force FM developed by the active element (the muscle) is calculated as the product of three factors: the muscle activation a( t), the force-length factor f ( / ), and the force-speed factor ( / ) (or muscle force indicator):

[0068] F^( / )

[0069] Muscle activation a(t) represents the activation level of the muscle between 0 and 1. It takes into account the ratio of recruited motor units, the discharge frequency of these motor units, the type of recruited fibers, and the sarcomeric dynamics within the muscle. a(t) is estimated as a function of the nerve excitation level e(t) of the muscle with a nonlinear time equation, and a typical nonlinear static force-EMG relationship, obtained from the literature. Other more physiological models using the mass principle could also be considered depending on the conditions of the species.

[0070] The force-length (FL) relationship provides the maximum force that can be produced a muscle as a function of its length under isometric conditions. While it originates at the molecular level and describes the average overlap of sarcomeric filaments within the muscle, this FL relationship has been measured in humans at both the molecular and muscle levels, where it takes a Gaussian form. Maximum force is obtained at a specific muscle length, called optimal isometric muscle length, which is specific to the muscle being studied. Charts are available for optimal isometric muscle length values ​​(see for example [Rajagopal et al., 2016]). During submaximal contractions, this FL relationship changes with the activation threshold a(t), as shown in dotted lines in [Fig. 4b]. In this figure, the solid curve represents a maximum activation level, and the dotted / broken curves represent lower activation.

[0071] The force-velocity (FV) relationship f (t) provides the instantaneous maximum force that a muscle can produce as a function of its contraction velocity, at optimal length and maximum activation level. It is assumed that fFV decreases and increases hyperbolically during concentric and eccentric efforts respectively. This relationship is therefore defined by two hyperbolas with a junction C2 at the point of zero contraction velocity. In this FV relationship, the maximum shortening velocity vmax and the rate of increase of the hyperbola increase and decrease respectively with the muscle activation level a, as shown in Figure 4a. Figure 4a shows the force-velocity fa) as a function of relative velocity L for different activation levels a and length 1. Figure 4a includes plots for slow fibers and plots for fast fibers.The solid line curve considers a maximum activation level while the dotted / broken line curves reflect the effect of the activation level. Figure 4b represents the force-length f(J, a) as a function of length 1 for different activation levels a.

[0072] The time evolution of muscle activation can be directly estimated from the nervous excitation e( t) with the neurophysiological models defined previously. In this case, a ( t, e ( t ) ) is obtained.

[0073] To estimate the temporal evolution of the length of the muscle [M ( t ) and its contraction speed yM ( f ), it is necessary to solve at each instant the problem of balance between the muscle and the tendon.

[0074] In all cases, / 7 ( / 1) = lMT(t) where lM1 (t) is obtained as pre previously indicated and lM(t) is the unknown variable. It follows by muscle-tendon equilibrium, and according to the active properties of the muscle that:

[0075] / "( / ), / "( / ) ]= / '[ / , a(?e(f))] = a(l))

[0076] Rearranging this expression, inverting the FV property, and observing that the muscle contraction velocity yM(?) is the derivative of the muscle length lM(t ) • H follows:

[0077] .wn -- r 1 f \ V) dt hv\a(t4t^

[0078] This differential equation takes as input the quantities lMT (j} and g(?) which are known (see previously), and has as its only variable the quantity / M(?) (researched). This equation is solved numerically at each time step for the variable [M( t ), from which the muscle contraction speed yM ( ? ) is immediately obtained by time differentiation.

[0079] Furthermore, depending on the implementations, it is possible to simplify this approach by setting the tendon length to a fixed constant f and assuming that the tendon transmits muscle force regardless of contraction conditions. In this case, lM and vM are directly known, as lM(t) = lMT(?) - lT and.— df1 _ dlMT , and the above differential equation ' u ' “ dt “ dt can be avoided (as can the calculations it generates, thus simplifying implementation and accelerating the calculation speed). This simplification can be carried out for certain muscles whose tendon is short or rigid (which is the case for most muscles in the human body), for example, or when the calculation resources are limited on the determination device.

[0080] Thus, using the neuromuscular and musculotendinous models, it is possible to estimate the three variables for determining the training load: a ( t ), lM (?) and vM ( ? ) from the temporal evolutions of the quantities of nervous excitation e(?) and musculotendinous lengths lMT( ?) obtained from the EMG and IMUs data.

[0081] These variables are then combined, as previously explained, to obtain the muscular constraints indicator sM(?) and the internal training load

[0082] The Hill neuromusculotendinous model defined previously is generic, since the muscle and tendon lengths lM and f used previously are normalized / relative. It is possible to scale this neuromusculotendinous model to a muscle-specific scale by integrating absolute values ​​for the lengths and defined previously. These values ​​are muscle-specific and are critical factors in the accuracy of muscle dynamics predictions based on sensitivity studies ([Scovil & Ronsky, 2006]; [Redl et al., 2007]; [Xiao & Higginson, 2010]; [Carbone et al., 2016]). These lengths and £ have been estimated in the past from cadaveric data ([Ward et al., 2009]). Thus, each muscle has its own ratio, and typical values ​​for an adult human of T 170cm are available in abacuses.

[0083] Other parameters in the neuromusculotendinous model can be adapted for each muscle, such as the rate of increase of the force-velocity relationship, or the slope of the EMG-force relationship. However, these values ​​are not available for all muscles in the human body, do not have a strong impact on muscle length predictions [M ( t ], and are therefore most often generic to the muscles studied. It is possible to calibrate these values ​​by minimizing the difference between experimental and estimated moments. Other parameters, such as tendon stiffness, can be estimated experimentally.

[0084] Finally, the maximum isometric force that a muscle can develop is different from one muscle to another, but the quantity 'absolute force' is not useful in the approach developed here.

[0085] Furthermore, the MSQ model in Figure 3 and used to obtain the musculotendinous lengths lMT from the kinematic data / (;) is also generic, and does not represent the bone geometry of the subject studied, who has specific bone segment lengths. This generic MSQ model may therefore be limited in describing the musculotendinous lengths jMT which vary non-linearly with the bone lengths. It is possible to scale the generic MSQ model by giving it the same size as the individual studied. In this case, all the lengths of the skeletal segments are linearly adjusted, and multiplied by a factor k, which is the ratio of the individual's height to that of the generic model (individual of height 170cm). The non-linear adjustment of the musculotendinous lengths lMT can then be recalculated with the scaled MSQ model.This scaling can be done automatically, for example, with the OpenSim simulation platform. This scaling approach remains limited, however, since the bone length ratios remain constant, before and after scaling, and specific to the generic MSQ model and not to the individual. A higher degree of specificity can be obtained by measuring the individual's segments, using non-linear bone length relationships from the literature, or by automatically creating subject-specific MSQ models from bone and muscle volumes obtained after segmentation of medical images of that subject. There are tools allowing the creation of such models (see for example [Modenese & Renault, 2021]). On the other hand, this . This approach requires medical images and a significant amount of time to segment the medical images.

[0086] In all cases, the musculotendinous lengths and f obtained experimentally for a generic human ([Ward et al., 2009]) are adjusted either linearly by the ratio of the lengths [MT before and after skeletal scaling, or with more advanced calibration methods ([Modenese et al., 2016]).

[0087] Other features and benefits

[0088] In an exemplary embodiment, the method can be implemented within the following overall framework.

[0089] A user is equipped with a smart garment comprising the sensors mentioned above, such as, for example, cycling shorts or leggings. According to the instructions mentioned by a third party (for example, a sports trainer, a coach, etc.), or according to a plan established by the user, the user performs one or more exercises involving all or part of the muscle and tendon groups covered by the smart garment. During the exercise, the smart garment captures data from the muscle group(s) being used. The data is recorded in a suitable recording device. This recording device may, depending on the embodiments, either be located in the smart garment itself or in a third-party device connected to the smart garment, such as, for example, a tablet or a smartphone, connected to the smart garment via a wired or wireless connection.The wired connection can be provided by a USB cable connected from the smartphone or tablet to the smart garment. This possibility is interesting in that it allows the smart garment not to need to be equipped with its own power supply, this being provided via the smartphone or tablet. This therefore makes the garment simpler to manufacture. The wireless connection can be in the form of a BLE™ (from the English for “Bluetooth Low Energy”) or Zigbee™ connection. This solution makes the manufacture of the smart garment more complex, but allows for greater flexibility in the use of the smart garment.

[0090] In any event, the data from the physical exercise performed by the user are stored as it is performed. Once stored, this data can be processed to determine the internal muscular load of the exercise performed by the user, according to the method for determining the internal muscular load as presented previously.

[0091] An example of a system for implementing all or part of the procedures described herein is described in relation to [Fig.6].

[0092] In this example, as explained previously, the user is equipped with a smart garment (VeTI), having BLE type wireless connectivity. The smart clothing (VeTI), is paired with a communication device (DisPC) of the tablet type, via this wireless connection. The communication device (DisPC) is connected to a communication network, for example via wireless connectivity of the WiFi or 4 / 5 G type. The communication device (DisPC) also includes an application, installed prior to the use of the smart clothing (VeTI), allowing: the storage of data captured during exercise from the smart clothing (VeTI); the transmission of all or part of this data to a processing server (SrVT); the reception, from this server (SrvT), of data representative of the internal muscular load. In this example, a remote server is therefore in charge of calculating the muscular load on behalf of the communication device (DisPC).Depending on the exemplary embodiments, the communication device (DisPC) can transmit all of the data recorded for a given exercise. In another exemplary embodiment, the communication device (DisPC) can also transmit pre-processed data: once the exercise data is recorded, the communication device (DisPC) can in fact clean and normalize this data in order, on the one hand, to facilitate the processing carried out by the processing server and, on the other hand, to limit the weight of the data to be transmitted to the server, on the one hand with the aim of reducing the bandwidth required for transmitting this data and, on the other hand, generally saving resources. The transmitted data can also be compressed to further reduce the volume to be transmitted.The processing server (SrVT), for its part, receives the data from the communication device (DisPC), and possibly decompresses them. Then it identifies, based on the content of the request transmitted to it by the communication device (DisPC), the physiological and morphological data of the user to whom these exercise recordings are associated as well as the muscle groups used during the exercise. The request transmitted by the communication device (DisPC) may include these pieces of information. The request transmitted by the communication device (DisPC) may also or alternatively include a user identifier. This user identifier is then extracted by the processing server to retrieve, within a database accessible by the server, the user's individualization characteristics (e.g. physiological and morphological data).

[0093] Based on this data, the server implements the method previously described to calculate an internal load of one or more muscles or muscle groups depending on the exercise performed whose recordings have been transmitted. Once this internal muscular load has been calculated, the server returns it to the communication device (DisPC) so that it can process it according to the instructions of application program installed on the communication device (DisPC). This application can process the internal muscular load calculated by the server by integrating it, for example, into an internal muscular load history or a cumulative index. This data can then be integrated into a training planning application module, so that this module can, based on the calculated internal muscular load, plan one or more additional training sessions aimed at reducing or, on the contrary, increasing this internal muscular load.

[0094] In relation to [Fig. 5], a simplified architecture of a device for determining internal muscle load capable of implementing the method for determining internal muscle load as presented previously is presented. Such a device for determining internal muscle load comprises a memory 51, a processing unit 52 equipped for example with a microprocessor, and controlled by the computer program 53, implementing the method according to the invention. In at least one embodiment, the invention is implemented in the form of an application installed on a communication device via a device dedicated solely to determining internal muscle load. Such a device for determining internal muscle load comprises for example all or part of the following means:

[0095] Means for obtaining kinematic data i(t) and electromyographic data emg(t) collected during a static or dynamic effort of said at least one muscle;

[0096] Means for determining, for said at least one muscle, from the kinematic data i(t) and electromyographic data emg(t), the musculotendinous length [Mr , a data item representative of the activation level a( t ), a data item representative of the muscle length lM(t), a data item representative of the contraction speed vM ( t ), and a force indicator Fpy ( t ) depending solely on the contraction speed vM(t);

[0097] Calculation means, configured to calculate an index of muscle constraints SM(t) from the muscle activation data a(t), the muscle length 1M(t) and the force indicator Fpy(t), and to calculate data representative of the internal muscle load CM from the muscle constraints sM(t) and a duration t of said, the data representative of the internal muscle load CM-

[0098] These means are in the form of a specific software application, or in the form of dedicated hardware components built for the purpose of performing these functions, such as a security element (SE) or a secure execution environment. The security element may be in the form of a Sim card, USim, UICC, or even a specific security component. Bibliographic references

[0099] [Rajagopal et al., 2016] : Rajagopal, Apoorva, et al. "Full-body musculoskeletal model for muscle-driven simulation of human gait." IEEE transactions on biomédical engineering 63.10 (2016): 2068-2079. [Modenese & Renault, 2021] : Modenese, Luca, and Jean-Baptiste Renault. "Automatic génération of personalised skeletal models of the lower limb from three-dimensional bone geometries." Journal of Biomechanics 116 (2021): 110186. [Scovil & Ronsky, 2006] : Scovil, Carol Y., and Janet L. Ronsky. "Sensitivity of a Hill-based muscle model to perturbations in model parameters." Journal of biomechanics 39.11 (2006): 2055-2063 [Redl et al., 2007] : Redl, Christian, Margit Gfoehler, and Marcus G. Pandy. "Sensitivity of muscle force estimâtes to variations in muscle-tendon properties." Human movement science 26.2 (2007): 306-319. [Xiao & Higginson, 2010] : Xiao, Ming, and Jill Higginson. "Sensitivity of estimated muscle force in forward simulation of normal walking." Journal ofapplied biomechanics 26.2 (2010): 142-149. [Carbone et al., 2016] : Carbone, Vincenzo, et al. "Sensitivity of subject-spécifie models to Hill muscle-tendon model parameters in simulations of gait." Journal of biomechanics 49.9 (2016): 1953-1960. [Ward et al., 2009] : Ward, Samuel R., et al. "Are current measurements of lower extremity muscle architecture accurate?." Clinical orthopaedics and related research 467.4 (2009): 1074-1082. [Modenese et al., 2016] : Modenese, Luca, et al. "Estimation of musculotendon parameters for scaled and subject spécifie musculoskeletal models using an optimization technique." Journal of biomechanics 49.2 (2016): 141-148. [Al Borno et al., 2022] : Al Borno, M., O’Day, J., Ibarra, V. et al. OpenSense: An open-source toolbox for inertial-measurement-unit-based measurement of lower extremity kinematics over long durations. J NeuroEngineering Rehabil 19, 22 (2022). https: / / doi.org / 10.1186 / sl2984-022-01001-x

Claims

Claims

1. Method for determining data representative of an internal muscular load (CM) developed by at least one muscle belonging to a muscular group, method implemented by means of an electronic device comprising a processor and a memory, said method comprising the following steps: - obtaining (A10) kinematic data (i(t)) and electromyographic data (emg(t)) collected during a contraction of said at least one muscle; - determination (A20), for said at least one muscle, from the kinematic data (i(t)), of data representative of a musculotendinous length QMT( / using a musculoskeletal model; - determination (A30), for said at least one muscle, from the electromyographic data (emg(l)), of data representative of a level of muscular activation (a(t)); - determination (A40), for said at least one muscle, from the musculotendinous length / Q), the muscle activation level (a(t)) and two neuromuscular and musculotendinous models (Ml, M2), a muscle length (ZV / ( / )) and data representative of a contraction speed (yM; - determination (A50), for said at least one muscle, of an estimate of a muscle deformation (dM(t)), from the muscle length (lM ( t )); - calculation (A60) of a muscle stress index (from muscle activation data (a(t)), muscle deformation (dM(t)) and a muscle strength indicator (fFV(t)) dependent on the contraction speed (AoP1 - calculation (A70) of the internal muscular load by the product of the muscular constraint index (sM( t) ) and a duration (t ) of said effort.

2. The method of claim 1, wherein the determining step (A20) of the musculotendinous length of said at least one muscle includes the sub-stages of: simulation of joint angles (“(t)) of joints set in motion during the effort, as a function of the kinematic data t)) and the musculoskeletal geometric model; and calculation, from the joint angles ('( / )) and the musculoskeletal geometry described by a musculoskeletal model of the temporal evolution of the musculotendinous length (jMT ( i \ ).

3.

4. A method according to claim 1 or claim 2, wherein the step of determining (A30) the data representative of the level of muscular activation (a(t)) of said at least one muscle comprises the sub-steps of: filtering, in the frequency domain, of electromyographic data (emg(t)) delivering signal envelopes (emg(t)); rectification of signal envelopes (emg(t)); normalization of signal envelopes (emg(t)), delivering normalized envelopes describing the level of nervous excitation t)) of the muscle; and calculation, from the level of nervous excitation (e( t )) of the muscle, of the data representative of the muscular activation («(0)- Electronic device, comprising a processor and a memory arranged for determining data representative of a load internal muscular developed by at least one muscle belonging to a muscle group, the device comprising: means for obtaining kinematic data (i(t)) and electromyographic data (emg(t)) collected during a static or dynamic effort of said at least one muscle; means of determining, for said at least one muscle, from kinematic (i(t)) and electromyographic (emg(l)y) data of a musculotendinous length of a data representative of a level of muscular activation a(t), of data representative of a muscular length of an estimation of a muscular deformation (dM (t)), and a data representative of a contraction speed (yM ( ; and calculation means, configured to calculate a muscle stress index (t)) from muscle activation data (¢ / ( / )), muscle deformation data (d M(t)) and a muscle strength indicator (fFV(t)) dependent on the contraction speed (y^( r)), and to calculate data representative of the internal muscle load (CM\

5. Electronic system comprising at least one processor and at least one memory arranged for the determination of data representative of an internal muscular load developed by at least one muscle belonging to a muscle group, said system comprising a smart garment (VeTI), paired with a communication device (DisPC) connected to a processing server (SrVT) through a communication network, said electronic system further comprising: - means for obtaining kinematic data (i(t)) and electromyographic data (emg(t)) collected during a static or dynamic effort of said at least one muscle in contact with the smart garment (VeTI);- means for determining, for said at least one muscle, from the kinematic (i(t)) and electromyographic (emg(t)) data, a musculotendinous length of a data item representative of an activation level (a(t)), of a data item representative of a muscle length (^(f)), of an estimation of a muscle deformation (dM(t)) and of a data item representative of a contraction speed; and; - calculation means, configured to calculate a muscle stress index (^( f)) from muscle activation data (( / )). muscle deformation (dM(t)) and a force indicator (fFV(t)) dependent on the contraction speed (yM( f)), and to calculate data representative of the internal muscle load (CMy - the means of obtaining, the means of determining and the means of calculating being distributed between the different components of the system.

6. A computer program product comprising program code instructions for implementing a determination method according to one of claims 1 to 3, when executed by a processor.

Images