Muscle force prediction method under functional electrical stimulation

Abstract

The invention provides a muscle force prediction method under functional electrical stimulation, which can be applied to the technical field of electrical stimulation. The muscle force prediction method comprises the following steps: acquiring muscle fiber length information and muscle fiber speed information of a target object; processing the electrical stimulation parameter by using a nerve activation model to obtain a nerve activation value; the nerve activation model represents an association relationship between a predetermined electrical stimulation parameter and a predetermined nerve activation value; the nerve activation value represents the nerve activation degree of the electrical stimulation parameter to the nerve; processing the nerve activation value by using a muscle activation model to obtain a muscle activation value; the muscle activation model represents an association relationship between a predetermined nerve activation value and a predetermined muscle activation value; the muscle activation value represents the activation degree of the target muscle under the condition that the nerve of the target muscle is in the nerve activation degree; and processing the muscle fiber length information, the muscle fiber speed information and the muscle activation value by using a muscle force prediction model, and predicting the muscle force information of the target muscle.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical stimulation, and in particular to a method for predicting muscle force under functional electrical stimulation. Background Art

[0002] Functional Electrical Stimulation (FES) applies electrical current to muscles, stimulating them to contract and enabling the user to perform movements autonomously. FES can improve muscle function, increase strength, and prevent muscle atrophy.

[0003] Obtaining accurate muscle force signals is crucial in adaptive control systems for functional electrical stimulation. However, delays in the process from electromyography to electrical stimulation nerve activation to muscle contraction, as well as the effects of stimulation artifacts, result in low accuracy in predicting limb-related information, making it difficult to achieve timely and accurate functional electrical stimulation. Summary of the Invention

[0004] In view of the above problems, the present invention provides a method for predicting muscle force under functional electrical stimulation.

[0005] According to one aspect of the present invention, a method for predicting muscle force under functional electrical stimulation is provided, comprising: obtaining muscle fiber length information and muscle fiber velocity information of a target object; the muscle fiber length information and muscle fiber velocity information respectively characterize the muscle fiber length and muscle fiber velocity of the target muscle when the target muscle of the target object is electrically stimulated according to electrical stimulation parameters and the target object performs an action task; processing the electrical stimulation parameters using a neural activation model to obtain a neural activation value; the neural activation model characterizes the correlation between predetermined electrical stimulation parameters and predetermined neural activation values; the neural activation value characterizes the degree of neural activation of the nerves by the electrical stimulation parameters; processing the neural activation value using a muscle activation model to obtain a muscle activation value; the muscle activation model characterizes the correlation between predetermined neural activation values ​​and predetermined muscle activation values; the muscle activation value characterizes the degree of activation of the target muscle when the nerves of the target muscle are at the neural activation degree; processing the muscle fiber length information, muscle fiber velocity information and muscle activation value using a muscle force prediction model to predict the muscle force information of the target muscle.

[0006] According to the embodiment of the present application, the present application directly establishes the relationship between the electrical stimulation parameters and the muscle activation. On this basis, by processing the electrical stimulation parameters using the neural activation model, the corresponding neural activation value can be determined in the case of stimulating the target muscle according to the electrical stimulation parameters. Then, the muscle activation model can be used to process the neural activation value to determine the corresponding muscle activation value. In this way, by directly processing the muscle fiber length information, muscle fiber velocity information and muscle activation value under the above electrical stimulation parameters using the muscle force prediction model to predict the muscle force information of the target object when performing the target task under the electrical stimulation of the above electrical stimulation parameters, the stimulation artifacts of electromyographic information or other physiological signals can be avoided to interfere with the predicted muscle force information, the prediction accuracy is improved, and the delay in the process of electromyographic information to electrical stimulation neural activation, and then to muscle contraction is avoided, and the timeliness of predicting the muscle force information is improved. BRIEF DESCRIPTION OF DRAWINGS

[0007] The above and other objects, features and advantages of the present application will become more apparent from the following description of embodiments of the present application taken in conjunction with the accompanying drawings, in which:

[0008] Figure 1 An application scenario diagram of the muscle force prediction method under functional electrical stimulation according to the embodiment of the present application is shown.

[0009] Figure 2 A flowchart of the muscle force prediction method under functional electrical stimulation according to the embodiment of the present application is shown.

[0010] Figure 3 A schematic diagram of the muscle force information acquisition method according to the embodiment of the present application is shown.

[0011] Figure 4 A schematic diagram of the predetermined relationship parameters according to the embodiment of the present application is shown.

[0012] Figure 5 A flowchart of the method for determining the second predetermined constant, the predetermined relationship parameters and the predetermined critical fusion frequency according to the embodiment of the present application is shown.

[0013] Figure 6 A mechanical property diagram of the simulated muscle-bone model according to the embodiment of the present application is shown.

[0014] Figure 7 A model diagram of the tendon unit according to the embodiment of the present application is shown.

[0015] Figure 8 A diagram of the composition of the active mechanics and contraction dynamics of the simulated muscle-bone model according to the embodiment of the present application is shown.

[0016] Figure 9A flow chart of a method for determining joint torque information according to an embodiment of the present invention is shown.

[0017] Figure 10 A block diagram of an electronic device suitable for implementing a muscle force prediction method under functional electrical stimulation according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0018] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present invention. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of embodiments of the present invention. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concept of the present invention.

[0019] The terms used herein are only for describing specific embodiments and are not intended to limit the present invention. The terms "comprise", "include", etc. used herein indicate the presence of the features, steps, operations and / or components, but do not exclude the presence or addition of one or more other features, steps, operations or components.

[0020] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0021] When expressions such as "at least one of A, B, and C, etc." are used, they should generally be interpreted in accordance with the meaning commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0022] In the technical solution of the present invention, the user information involved (including but not limited to user personal information, user image information, user device information, such as location information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data comply with relevant laws, regulations and standards, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0023] In scenarios where personal information is used for automated decision-making, the methods, devices, and systems provided by embodiments of the present invention provide users with corresponding operational portals, allowing them to choose to agree or reject the automated decision-making results; if the user chooses to reject, the expert decision-making process will be entered. The term "automated decision-making" herein refers to the activity of automatically analyzing and evaluating an individual's behavioral habits, interests, or economic, health, or credit status through computer programs and making decisions. The term "expert decision-making" herein refers to the activity of decision-making by individuals who specialize in a particular field, possess specialized experience, knowledge, and skills, and have reached a certain level of professional expertise.

[0024] Embodiments of the present invention provide a method for predicting muscle force under functional electrical stimulation, enabling direct prediction of muscle force using electrical stimulation parameters. Compared to some approaches that rely on musculoskeletal models established through electromyography (EMG) to predict muscle force, the present invention directly predicts muscle force using functional electrical stimulation parameters, reducing artifacts in EMG acquisition and interference from other physiological signals, improving prediction accuracy, and increasing the efficiency of muscle force prediction.

[0025] Figure 1 A diagram showing an application scenario of a muscle force prediction method under functional electrical stimulation according to an embodiment of the present invention is shown.

[0026] like Figure 1 As shown, the application scenario 100 according to this embodiment may include an electronic device 110 and a limb 120 of a target object. The electronic device 110 may be an electrical stimulation device, such as a functional electrical stimulator. For example, the electronic device 110 may be electrically connected to the limb 120 of the target object and electrically stimulate the muscles of the limb. In this case, the electronic device 110 may predict the muscle force of the target object based on the muscle fiber length information, the muscle fiber velocity information and the electrical stimulation parameters. Based on this, the muscle force prediction method under functional electrical stimulation provided in the embodiment of the present invention can generally be performed by the electronic device 110. It should be understood that Figure 1 The number of electronic devices in the embodiment is merely illustrative. Any number of electronic devices may be provided according to implementation requirements.

[0027] The following will be based on Figure 1 The scene described by Figures 2 to 9 The muscle force prediction method under functional electrical stimulation according to an embodiment of the invention is described in detail.

[0028] Figure 2 A flow chart of a method for predicting muscle force under functional electrical stimulation according to an embodiment of the present invention is shown.

[0029] like Figure 2As shown, the muscle force prediction method under functional electrical stimulation of this embodiment includes operations S210 to S240.

[0030] In operation S210 , muscle fiber length information and muscle fiber velocity information of the target object are acquired.

[0031] In operation S220 , the electrical stimulation parameters are processed using the neural activation model to obtain a neural activation value.

[0032] In operation S230 , the nerve activation value is processed using a muscle activation model to obtain a muscle activation value.

[0033] In operation S240 , the muscle fiber length information, the muscle fiber velocity information, and the muscle activation value are processed using a muscle force prediction model to predict muscle force information of the target muscle.

[0034] It should be noted that, in embodiments of the present invention, user consent or authorization may be obtained before obtaining user information. For example, before executing various operations in embodiments of the present invention, a request to obtain user information may be issued to the user. If the user consents or authorizes obtaining user information, the various operations in embodiments of the present invention are executed.

[0035] In embodiments of the present invention, a corresponding operation portal can be provided for the user to choose to agree or reject the automated decision result. That is, before processing / decision-making on the user's information, an instruction to agree or reject the processing / decision-making process can be obtained from the user through the corresponding operation portal. If the user agrees to the processing / decision-making process, the processing / decision-making process is carried out on the user's information, that is, the various operations of the embodiments of the present invention are executed. If the user rejects the processing / decision-making process, the expert decision-making process is entered.

[0036] In addition, the technical solution of the present invention is only used to predict the muscle force information of the target muscle and is not used for diagnosis and treatment of diseases.

[0037] Specifically, in embodiments of the present invention, the target subject may be a user whose muscle force information is to be predicted. In one embodiment, electrodes may be placed on the skin surface of the target subject's target muscle, and electrical stimulation parameters may be input into an electronic device, so that the electronic device electrically stimulates the target subject's target muscle via the electrodes. For example, the target muscle may be a single muscle unit. The electrical stimulation parameters may include, but are not limited to, electrical stimulation intensity and electrical stimulation frequency.

[0038] At the same time, the target object may also be performing an action task. Action tasks may include static action tasks and dynamic action tasks. Static action tasks may include, but are not limited to, standing tasks, etc. Dynamic action tasks may include, but are not limited to, walking tasks, etc. It should be noted that the present invention is particularly suitable for tasks in which muscles do not undergo voluntary contraction under static electrical stimulation. Furthermore, in the present invention, the muscle fiber length information and muscle fiber velocity information of the target object may be obtained. The muscle fiber length information and the muscle fiber velocity information respectively characterize the muscle fiber length and muscle fiber velocity of the target muscle when the target muscle of the target object is electrically stimulated according to the electrical stimulation parameters and the target object performs an action task.

[0039] The electrical stimulation parameters can then be input into a neural activation model, which outputs a neural activation value. The neural activation model can be a preset model stored in a memory of an electronic device. The neural activation model represents the relationship between the electrical stimulation parameters and the neural activation value. The neural activation value represents the degree of neural activation caused by the electrical stimulation parameters, specifically the degree of neural activation caused by the electrical stimulation parameters on the target muscle.

[0040] The nerve activation value can then be input into a muscle activation model, which outputs a muscle activation value. The muscle activation model can also be a preset model stored in the memory of the electronic device. The muscle activation model represents the correlation between the nerve activation value and the muscle activation value. The muscle activation value represents the activation level of the target muscle, i.e., the muscle excitability, when the nerves of the target muscle are at the aforementioned nerve activation level.

[0041] Subsequently, the muscle fiber length information, muscle fiber velocity information, and muscle activation value can be input into a muscle force prediction model to obtain muscle force information of the target muscle. For example, the muscle force prediction model characterizes the correlation between muscle fiber length information, muscle fiber velocity information, muscle activation value, and muscle force information. The muscle force information characterizes the muscle force of the target muscle when stimulated according to the electrical stimulation parameters. For example, the neural activation model, muscle activation model, and muscle force prediction model can be mathematical models.

[0042] Based on the above content, the present invention directly establishes the relationship between electrical stimulation parameters and muscle activation. On this basis, by using a neural activation model to process the electrical stimulation parameters, the corresponding neural activation value when the target muscle is stimulated according to the electrical stimulation parameters can be determined. Then, the muscle activation model can be used to process the neural activation value to determine the corresponding muscle activation value. In this way, by using a muscle force prediction model to directly process the muscle fiber length information, muscle fiber velocity information and muscle activation value under the above-mentioned electrical stimulation parameters, the muscle force information of the target object when performing the target task under the electrical stimulation of the above-mentioned electrical stimulation parameters can be predicted. This can avoid the stimulation artifacts of the electromyographic information or other physiological signals from interfering with the predicted muscle force information, thereby improving the prediction accuracy, and avoiding the delay in the process from electromyographic information to electrical stimulation neural activation and then to muscle contraction, thereby improving the timeliness of predicting muscle force information.

[0043] In one embodiment of the present invention, muscle fiber length information and muscle fiber velocity information are determined based on the muscle-tendon unit length information and the tendon model. The muscle-tendon unit length information is collected when the target muscle of the target object is electrically stimulated according to the electrical stimulation parameters and the target object performs an action task, and the motion trajectory information and ground reaction force information are analyzed and determined based on the predetermined musculoskeletal model of the target object.

[0044] Figure 3 A schematic diagram of a method for acquiring muscle force information according to an embodiment of the present invention is shown.

[0045] like Figure 3As shown, data collection and preprocessing can first be performed. Specifically, when the target object is performing an action task, an optical motion capture system can be used to collect the target object's marker points. Furthermore, when the target object is performing the action task, ground reaction force information of the ground acting on the target object can be collected. Then, based on the motion trajectory information and ground reaction force information formed by the marker points, a simulated musculoskeletal model of the target object can be established in simulation software. For example, the simulation software can be general biomechanics simulation software, such as OpenSim simulation software. The simulated musculoskeletal model can be a geometric model, specifically a Hill-type muscle-tendon unit musculoskeletal model. This musculoskeletal model can then be analyzed using inverse kinematics (IK), inverse dynamics (ID), and muscle analysis in the simulation software to obtain information such as the target object's joint angle information, joint torque information, muscle-tendon unit length information, and muscle moment arm information. For example, the muscle-tendon unit length can be used to determine the muscle fiber length information, but the embodiments of the present invention are not limited to this. In other embodiments of the present invention, the above-mentioned information can also be obtained in combination with the data measured by the inertial measurement unit, which will not be elaborated here. Then, based on the muscle fiber length information and the exercise duration of the target muscle, the muscle fiber velocity information can be determined. However, it should be understood that this is only an example. In other embodiments, the muscle fiber length information and muscle fiber velocity information can also be obtained by other means, which will not be elaborated here. In this way, by using the neural activation model to process the electrical stimulation parameters, the neural activation value is obtained, and then the muscle activation model is used to process the neural activation value to obtain the muscle activation value. On this basis, the muscle activation degree, muscle fiber length and velocity under the electrical stimulation parameters can be combined by the muscle force prediction model to predict the muscle force information.

[0046] On this basis, when the target object performs different action tasks, or determines the muscle fiber length and muscle fiber velocity of different muscles, the collected muscle fiber length information and muscle fiber velocity information may also be different. In this way, based on the muscle fiber length information of multiple action tasks, the muscle fiber velocity information of multiple action tasks, the muscle activation values ​​of multiple action tasks, and the electrical stimulation parameters of multiple action tasks, a muscle force prediction model with different parameters for multiple action tasks can be constructed. On this basis, a mapping relationship between the target object's multiple action tasks, multiple muscles, and multiple muscle force prediction models can be constructed. In this way, based on the action task being performed by the target object and the information of the target muscle corresponding to the action task, the corresponding muscle force prediction model can be queried through this mapping relationship, and the muscle force prediction model can be used to process the target object's corresponding muscle fiber length information, muscle fiber velocity information, and muscle activation value to obtain muscle force information. In this way, the accuracy of the predicted muscle force information can be improved.

[0047] In another embodiment of the present invention, there may be multiple target muscles. Multiple target muscles are distributed around the target joint of the target object. On this basis, the above method may further include: obtaining muscle lever arm information of each target muscle. According to the predetermined weight of each target muscle, the muscle force information of each target muscle and the muscle lever arm information of each target muscle, the joint torque of the target joint is determined. The error value of the joint torque relative to the predetermined joint torque is determined. When the error value does not meet the predetermined condition, the electrical stimulation parameter is adjusted based on the error value to obtain the adjusted electrical stimulation parameter. For example, the predetermined condition may be a predetermined error threshold. When the error value is greater than or equal to the predetermined error threshold, it can be determined that the error value does not meet the predetermined condition. The present invention does not limit the specific value of the predetermined error threshold.

[0048] Then, the joint torque can be compared with a predetermined joint torque set in advance to obtain an error value of the joint torque relative to the predetermined joint torque. For example, the predetermined joint torque can be the joint torque of the joint in the target state. For example, the target state is a healthy state, but is not limited to this. The predetermined joint torque can be a joint torque set by the user, for example, it can be a joint torque expected by the user, and the present invention does not limit this. On this basis, the electrical stimulation parameters can be adjusted based on the error value so that the target joint torque of the target object can be closer to the predetermined joint torque, thereby reducing the error between the measured joint torque and the predetermined joint torque. Specifically, the adjustment direction of the electrical stimulation parameters can be determined according to the sign of the error value.

[0049] For example, the error value can be a positive value or a negative value. In the case of a positive error value, the electrical stimulation parameter (for example, the electrical stimulation intensity) can be adjusted by a predetermined value to obtain an adjusted electrical stimulation parameter; in the case of a negative error value, the electrical stimulation parameter can be adjusted by a predetermined value to obtain an adjusted electrical stimulation parameter.

[0050] And the embodiments of the present application are not limited to this, in other embodiments of the present application, the relationship between the experimental joint torque, angle, acceleration and the experimental electrical stimulation parameter under electrical stimulation can also be determined by calculating the kinetic energy and potential energy of the target joint in combination with the Lagrange modeling method. Then, the joint angle information corresponding to the joint torque can be determined according to the relationship. Then, the electrical stimulation curve conforming to the angle information is determined. Then, the electrical stimulation parameter of the functional electrical stimulator is set according to the electrical stimulation curve, so as to realize the adjustment of the electrical stimulation parameter, thereby realizing accurate electrical stimulation motion control.

[0051] In an embodiment of the present application, processing the electrical stimulation parameter by using the neural activation model to obtain the neural activation value can include: calculating a first scaling factor based on the electrical stimulation intensity and a predetermined threshold. Calculating a second scaling factor based on the electrical stimulation frequency, a first predetermined constant, a second predetermined constant, a predetermined frequency constant and a predetermined relationship parameter. The first predetermined constant is calculated based on the second predetermined constant, the predetermined frequency constant and the predetermined relationship parameter. The predetermined frequency constant is calculated based on the second predetermined constant, a predetermined critical fusion frequency and the predetermined relationship parameter. The predetermined critical fusion frequency represents a critical frequency of stimulating muscle contraction. The second predetermined constant is a ratio of the predetermined maximum contraction force of the target muscle to the contraction force of the target muscle due to electrical stimulation when the electrical stimulation frequency is the predetermined critical fusion frequency. The predetermined relationship parameter is used to represent the relationship between the predetermined electrical stimulation frequency and the muscle force. The first scaling factor and the second scaling factor are processed by using the neural activation model to obtain the neural activation value.

[0052] For example, in the present application, the scaling factor is determined according to the electrical stimulation intensity U represented by amplitude or pulse width and the electrical stimulation frequency f.

[0053] On this basis, the neural activation model is as follows:

[0054] (1)

[0055] Wherein, e(t) represents the neural activation value; S u represents the first scaling factor; S f represents the second scaling factor. It should be noted that the meaning of the symbols here is only used to explain the content of formula (1). Similarly, the meaning of the symbols in other formulas below is also only used to explain the content of the corresponding formula, which will not be described one by one hereinafter.

[0056] The first scaling factor is calculated as follows:

[0057] (2)

[0058] Among them, S u Indicates the first scale factor; U tr Indicates the first voltage threshold; U stim Indicates the intensity of electrical stimulation; U sat represents the second voltage threshold. When the electrical stimulation intensity is less than the first voltage threshold, almost no muscle fibers are recruited; and when the electrical stimulation intensity is greater than the second voltage threshold, almost all muscle fibers are recruited.

[0059] The second scaling factor is calculated as follows:

[0060] (3)

[0061] Among them, S f represents the second proportional factor; a1 represents the first predetermined constant; a2 represents the second predetermined constant, which can be determined by experiment; f stim represents the electrical stimulation frequency; f0 represents a predetermined frequency constant; and R represents a predetermined relationship parameter. Specifically, R can be a shape parameter used to describe the relationship between the electrical stimulation frequency and muscle force, reflecting the steepness of the curve.

[0062] Figure 4 A schematic diagram showing predetermined relationship parameters according to an embodiment of the present invention is shown.

[0063] like Figure 4 As shown, the predetermined relationship parameter is correlated with the critical fusion frequency.

[0064] On this basis, assuming that the muscle force f=0, S f =0, when the electrical stimulation frequency is the predetermined critical fusion frequency S f =1, then a1 and f0 can be obtained.

[0065] For example, the first predetermined constant is calculated by the following formula:

[0066] (4)

[0067] Wherein, a1 represents a first predetermined constant; a2 represents a second predetermined constant; f0 represents a predetermined frequency constant; and R represents a predetermined relationship parameter.

[0068] The predetermined frequency constant is calculated by the following formula:

[0069] (5)

[0070] Wherein, f0 represents a predetermined frequency constant; a1 represents a first predetermined constant; a2 represents a second predetermined constant; R represents a predetermined relationship parameter; f CF Represents the predetermined critical fusion frequency.

[0071] Then, the muscle activation model can be used to process the neural activation value to obtain the muscle activation value, including: using the muscle activation model to process the neural activation value to obtain the muscle activation value. The coefficients of the muscle activation model are determined based on a first predetermined time constant, a second predetermined time constant, and a third predetermined time constant. The first predetermined time constant represents a predetermined time constant for electrically stimulating the muscle; the second predetermined time constant represents a predetermined time constant for the time it takes for the muscle force to rise from a baseline value to a peak value after the muscle is activated. The third predetermined time constant represents a predetermined time constant for the time it takes for the muscle force to fall from a peak value to a baseline value after the muscle activation ends.

[0072] For example, the muscle activation model can be expressed as the following second-order ordinary differential equation:

[0073] (6)

[0074] (7)

[0075] (8)

[0076] Where a(t) represents the muscle activation value; represents the first-order differential of a(t); represents the second-order differential of a(t); e(t) represents the neural activation value; k1 represents the first predetermined coefficient; k2 represents the second predetermined coefficient; T e Represents the predetermined time constant when electrically stimulating the muscle; T rise / fall Can represent T rise or T fall ;T rise T represents a predetermined time constant used to characterize the time it takes for muscle force to rise from a baseline value to a peak value after muscle activation; fall T represents a predetermined time constant used to characterize the time it takes for muscle force to decrease from peak value to baseline value after muscle activation ends. e 、T rise 、T fall Depends on physiological cross-sectional area, muscle mass and percentage of fast-twitch fibers. Since physiological cross-sectional area, muscle mass and percentage of fast-twitch fibers are related to muscle atrophy, the T e 、T rise 、T fall Introducing a muscle activation model allows the model to be applied to subjects with muscle atrophy, thereby improving the accuracy of neural activation values ​​and, consequently, the accuracy of predicted muscle force information.

[0077] In one embodiment of the present invention, the model parameters of the neural activation model, the muscle activation model, and the muscle force prediction model are determined by the following method:

[0078] In the i-th iteration round, based on the i-th random perturbation, the i-1-th model parameter is updated to obtain candidate model parameters. For example, the i-th random perturbation can be determined based on the i-th parameter optimization step size and the i-th perturbation random value randomly generated in the i-th iteration round. In other embodiments of the present invention, the i-1-th model parameter can be adjusted (e.g., decreased or increased) according to the i-th random perturbation to obtain adjusted candidate model parameters.

[0079] Then, the i-th electrical stimulation parameter can be obtained, and the candidate muscle force information can be calculated based on the i-th electrical stimulation parameter and the candidate model parameters. For example, the candidate model parameters can be used as the initial neural activation model, the initial muscle activation model, and the initial muscle force prediction model parameters to process the i-th electrical stimulation parameter to obtain the candidate muscle force information. The i-th electrical stimulation parameter can be a parameter different from the i-1-th electrical stimulation parameter, can be manually preset, or can be randomly generated, and the present invention is not limited to this.

[0080] Then, the joint torque model can be used to process the candidate muscle force information to obtain the candidate joint torque. For example, the joint torque model can correspond to the following formula:

[0081] (9)

[0082] In formula (9), M represents the joint torque of the target joint; n represents the number of target muscles around the target joint, that is, the number of muscles across the target joint, which can be preset; γ q is the predetermined weight of the qth muscle, which can be predetermined by experiments and is not limited here; F mt,i is the muscle force information of the target muscle of the i-th block, which can correspond to the candidate muscle force information during training; r mt,i is the muscle force arm information of the i-th muscle. Wherein, n is a positive integer greater than 1, and q is a positive integer less than or equal to n. On this basis, the joint torque of the target joint of the target object can be obtained by formula (9). It should be noted that the joint torque model can also be applied to the embodiment of adjusting the electrical stimulation parameters described above, which will not be repeated here. Then, the first difference value of the candidate joint torque relative to the experimental joint torque can be determined. Wherein, the experimental joint torque represents the actual value of the joint torque measured by the target object during the walking test. For example, for each object, the same walking test can be used to calibrate the model parameters. At the same time, the optimization objective function f is setcal , that is, the difference between the predicted joint torque and the experimental joint torque. For example, the objective function value f cal Calculated by the following formula:

[0083] (10)

[0084] (11)

[0085] (12)

[0086] (13)

[0087] Among them, N trials Indicates the number of walking tasks performed by the target subject. N DOFs Represents the predetermined degrees of freedom, which is used to represent the number of adjustable parameters considered when fitting the model parameters. t,d Represents the error between the candidate joint torque and the experimental joint torque. t represents the tth walking task. d represents an adjustable parameter. N r N represents the number of tendon data points in each walking task. rows Indicates the number of rows of data points collected for each walking task of the target object. t,d,r represents the experimental joint torque of the r-th tendon data point. represents the candidate joint torque of the r-th tendon data point. is the variance of the data points for each walking task, which is used to normalize the differences and reduce the differences between the data points corresponding to the experimental joint torques and the data points corresponding to the candidate joint torques. r Represents a predetermined penalty function to prevent non-physiological results. MTUs is the number of muscle-tendon units. P(r,j) represents the penalty value for the r-th muscle-tendon data point and the j-th muscle-tendon unit. is the normalized length of the jth muscle-tendon unit. When the value of deviates from the static length by more than 0.5, the penalty value P(r,j) is , used to penalize non-physiological muscle tension parameters. Other circumstances except

[0088] Thus, when the first difference value is less than the second difference value, the candidate joint torque is determined as the i-th joint torque, the candidate model parameter is determined as the i-th model parameter, and the second difference value is the difference value of the i-1-th joint torque relative to the experimental joint torque; when the first difference value is greater than or equal to the second difference value, and the acceptance probability of the candidate joint torque is greater than the i-th random number randomly generated in the i-th iteration round, the candidate joint torque is determined as the i-th joint torque, and the candidate model parameter is determined as the i-th model parameter; when the first difference value is greater than or equal to the second difference value, and the acceptance probability of the candidate joint torque is less than or equal to the i-th random number, the i-1-th joint torque is determined as the i-th joint torque, and the i-1-th model parameter is determined as the i-th model parameter. For example, the following formula can be used to calculate the acceptance probability of the candidate joint torque based on the difference between the candidate joint torque and the i-1-th joint torque and the i-th annealing temperature:

[0089] (14)

[0090] Where P represents the probability of acceptance. j represents the candidate joint torque, E i-1 represents the i-1 joint torque; T represents the annealing temperature.

[0091] When the acceptance probability is greater than the random number, the candidate joint torque can also be determined as the i-th joint torque, otherwise, the i-1-th joint torque is determined as the i-th joint torque. As the temperature drops, the acceptance probability will decrease accordingly. Then, when the candidate joint torque is determined to be the i-th joint torque, the candidate model parameter is determined to be the i-th model parameter. Calibrating the model parameters of the above-mentioned neural activation model, muscle activation model and muscle force prediction model in the above-mentioned manner can improve the accuracy of the muscle activation value, thereby improving the accuracy of the predicted muscle force information. It should be supplemented that the random numbers of each iterative round can be the same or different, and the present invention does not limit this.

[0092] Afterwards, the i-th parameter optimization step length can be updated to obtain the i+1-th parameter optimization step length according to the predetermined step length adjustment strategy based on the i-th acceptance rate of the candidate joint torque and the predetermined acceptance rate threshold. For example, when the i-th acceptance rate is greater than the predetermined acceptance rate, the step length can be increased according to the predetermined step length value; when the i-th acceptance rate is less than the predetermined acceptance rate, the step length can be reduced according to the predetermined step length value. The i-th acceptance rate characterizes the proportion of the joint torque determined by the candidate joint torque in the i-iteration round to the i-joint torque. For example, in the i-iteration round, a total of i candidate joint torques can be determined, and among the i candidate joint torques, k candidate joint torques are determined as the joint torques of the next round, where k is a positive integer less than or equal to i. The i-th acceptance rate is the proportion of the k candidate joint torques to the i candidate joint torques. It should be understood that, relative to the i-th joint torque, the i+1-th joint torque is the joint torque of the next round. In this way, by updating the step size round by round, it can be ensured that the algorithm can sample the solution space widely and avoid falling into local optimality.

[0093] Thereafter, the i-th annealing temperature parameter may be updated based on the number of updates of the i+1-th parameter optimization step, to obtain the i+1-th annealing temperature parameter. For example, after determining the i+1-th parameter optimization step, the i-th annealing temperature parameter may be reduced according to a predetermined cooling strategy to obtain the i+1-th annealing temperature parameter. Specifically, the i-th annealing temperature parameter may be reduced according to a predetermined temperature decrease rate to obtain the i+1-th annealing temperature parameter.

[0094] When the i+1th annealing temperature parameter is less than or equal to the predetermined temperature threshold, or the difference between the first difference value and the second difference value is less than the predetermined difference threshold, or i+1=m, the i-th model parameter can be determined as the model parameter of the neural activation model, the muscle activation model and the muscle force prediction model, where m is the predetermined round threshold.

[0095] On this basis, the above method allows occasional acceptance of poor solutions by dynamically adjusting the search range and acceptance criteria, thereby escaping from local optimality. It is suitable for high-dimensional, nonlinear optimization problems such as muscle activation models.

[0096] The following is described with reference to specific embodiments:

[0097] First, the initial model parameters can be determined. For example, taking the second predetermined constant, predetermined relationship parameter, and predetermined critical fusion frequency as an example, an initial solution (or state) can be selected for the second predetermined constant, predetermined relationship parameter, and predetermined critical fusion frequency, namely, the first predetermined constant, the first relationship parameter, and the first critical fusion frequency. The initial muscle force information, namely, the first muscle force information, can then be calculated using the first joint torque, the first predetermined constant, the first relationship parameter, and the first critical fusion frequency corresponding to the first electrical stimulation parameter. It should be noted that the model parameters to be optimized are not limited to the second predetermined constant, the predetermined relationship parameter, and the predetermined critical fusion frequency, but may also include the first predetermined coefficient, the second predetermined coefficient, the second proportionality factor, and the like, which are not limited in the present invention. An initial temperature can be set; in the present invention, it can be set to 200,000. The first predetermined constant, the first relationship parameter, and the first critical fusion frequency can then be iteratively updated in successive rounds, with the temperature being lowered in accordance with the cooling strategy in each iteration until the second predetermined constant, the predetermined relationship parameter, and the predetermined critical fusion frequency are achieved.

[0098] For example, the i-th random perturbation r×V can be superimposed on the current parameters (e.g., the i-1th constant, the i-1th relationship parameter, and the i-1th critical fusion frequency), where r represents a random value, r∈[-1,1), and V is the parameter optimization step size. This results in candidate constants, candidate relationship parameters, and candidate critical fusion frequencies as new solutions. The candidate joint torques corresponding to these candidate constants, candidate relationship parameters, and candidate critical fusion frequencies are then determined. Based on this, the more optimal candidate joint torque or the i-1th joint torque can be determined. If the candidate joint torque is more optimal than the i-1th joint torque, the candidate joint torque is determined as the i-th joint torque. Otherwise, the Metropolis acceptance criterion is used to calculate the acceptance probability of the candidate joint torque based on the difference between the candidate joint torque and the i-1th joint torque and the current temperature. For example, the acceptance probability can be a Boltzmann probability. The acceptance probability decreases as the temperature decreases.

[0099] Then, the parameter optimization step can be adjusted according to a predetermined step adjustment strategy, so that the acceptance rate of the new solution is close to 50% (corresponding to the acceptance rate threshold described above): if the acceptance rate is too high, the parameter optimization step is increased to expand the search range; if it is too low, the parameter optimization step is reduced to refine the search. In the present application, the parameter optimization step can be adjusted once every predetermined number of iteration rounds (for example, 15 times), but it should be understood that the present application is not limited thereto, and the parameter optimization step can also be adjusted for each iteration round. Then, the temperature can be reduced according to a cooling strategy, which determines the speed of temperature reduction (i.e., the predetermined temperature reduction rate described above) and the value of the final temperature (i.e., the predetermined parameter threshold described above). In the present application, the temperature reduction rate is set to 0.1, and after performing step adjustment for 5 times, the temperature is attenuated at the predetermined temperature reduction rate to reduce the probability of accepting a poor solution, so as to gradually converge to the optimal solution, i.e., the optimal second predetermined constant, predetermined relationship parameter and predetermined critical fusion frequency.

[0100] wherein when the temperature is reduced to a predetermined temperature threshold, or the difference between the first difference value and the second difference value is less than a predetermined difference threshold, or the iteration round is the mth iteration round, the algorithm ends, and in the present application, m can be set to 2000000 and the predetermined difference threshold can be set to 0.001.

[0101] Figure 5 A flowchart of a method for determining a second predetermined constant, a predetermined relationship parameter and a predetermined critical fusion frequency according to an embodiment of the present application is shown.

[0102] As shown in Figure 5 the method of this embodiment includes operations S501-S511.

[0103] In operation S501, an initial solution is determined.

[0104] In operation S502, a new solution is randomly generated.

[0105] In operation S503, it is determined whether the objective function value of the new solution is less than the objective function value of the current solution. If not, operation S504 is performed; if yes, operation S506 is performed.

[0106] In operation S504, the acceptance probability of the new solution is calculated.

[0107] In operation S505, it is determined whether the acceptance probability is greater than a random number. If not, operation S507 is performed; if yes, operation S506 is performed.

[0108] In operation S506, the new solution is accepted.

[0109] In operation S507, the current solution is maintained.

[0110] In operation S508, a new solution is determined.

[0111] In operation S509 , the annealing temperature parameter is reduced according to the cooling strategy, and the parameter optimization step size is updated according to the predetermined step size adjustment strategy based on the acceptance rate of the new solution and the predetermined acceptance rate threshold.

[0112] In operation S510, it is determined whether the reduced annealing temperature parameter is less than or equal to a predetermined temperature threshold, the difference between the first difference value and the second difference value is less than a predetermined difference threshold, or whether the number of iterations is greater than or equal to a predetermined number of iterations. If not, the process returns to operation S502; if so, the process proceeds to operation S511.

[0113] In operation S511 , model parameters are determined.

[0114] On this basis, different Hill-type muscle-tendon unit musculoskeletal models can be selected, and parameters can be fitted by the above method to obtain different muscle activation models. In this way, the muscle fiber length information and muscle fiber velocity information can be determined specifically according to different target objects using the corresponding Hill-type muscle-tendon unit musculoskeletal model, and the muscle activation value can be determined using the corresponding muscle activation model. Then, the muscle fiber length information, muscle fiber velocity information and muscle activation value are processed using the corresponding muscle force prediction model to obtain accurate muscle force information, which provides a basis for the safe, stable and precise control of the functional electrical stimulation system. Moreover, the corresponding electrical stimulation parameters can be determined based on the muscle force information to perform accurate electrical stimulation on the target object, so that functional electrical stimulation can be achieved in a timely and accurate manner.

[0115] In one embodiment of the present invention, the muscle fiber length information may include instantaneous muscle fiber length information and optimal muscle fiber length information. The muscle fiber velocity information may include instantaneous muscle fiber velocity information and optimal muscle fiber velocity information. The muscle fiber length information, muscle fiber velocity information and muscle activation value are processed by a muscle force prediction model to predict the muscle force information of the target muscle, including: normalizing the instantaneous muscle fiber length information using the optimal muscle fiber length information to obtain normalized target muscle fiber length information. For example, the ratio of the instantaneous muscle fiber length information to the optimal muscle fiber length information can be determined, and the ratio can be used as the normalized target muscle fiber length information. The optimal muscle fiber length information can refer to the optimal length of the muscle fibers of the target muscle when the muscle excitability is strongest. The instantaneous muscle fiber length information can be the muscle fiber length corresponding to the muscle activation value of the target muscle, that is, the instantaneous muscle fiber length of the target muscle.

[0116] Then, the optimal muscle fiber velocity information can be used to normalize the instantaneous muscle fiber velocity information to obtain normalized target muscle fiber velocity information. For example, the ratio of the instantaneous muscle fiber velocity information to the optimal muscle fiber velocity information can be determined, and the ratio can be used as the normalized target muscle fiber velocity information. Among them, the optimal muscle fiber velocity information can refer to the optimal velocity of the muscle fibers of the target muscle when the muscle excitability is strongest, for example, it can be obtained by differentiating the above-mentioned optimal muscle fiber length information. The instantaneous muscle fiber velocity information can be the muscle fiber velocity corresponding to the muscle activation value of the target muscle, that is, the instantaneous muscle fiber velocity of the target muscle, for example, it can be obtained by differentiating the above-mentioned instantaneous muscle fiber length information.

[0117] The muscle force prediction model can then be used to process the target muscle fiber length information, target muscle fiber velocity information, and muscle activation value to obtain muscle force information. For example, the target muscle fiber length information, target muscle fiber velocity information, and muscle activation value can be input into the muscle force prediction model to obtain muscle force information.

[0118] Specifically, the muscle force prediction model may include an active contraction force prediction model, a passive contraction force prediction model, and a muscle force prediction sub-model. Based on this, the muscle force prediction model is used to process muscle fiber length information, muscle fiber velocity information, and muscle activation value to predict the muscle force information of the target muscle, including: using the active contraction force prediction model to process the target muscle fiber length information, target muscle fiber velocity information, and muscle activation value to obtain target active contraction force information of the target muscle. The active contraction force prediction model characterizes the correlation between the predetermined muscle activation value, muscle fiber length, muscle fiber velocity, and the active contraction force of the actively contracting muscle units in the muscle. Subsequently, the passive contraction force prediction model can be used to process the target muscle fiber length information to obtain target passive contraction force information of the target muscle. The passive contraction force prediction model characterizes the correlation between the predetermined muscle activation value, muscle fiber length, and the passive contraction force of the passively contracting muscle units in the muscle. Furthermore, the muscle force prediction sub-model can be used to process the target active contraction force information and the target passive contraction force information to obtain muscle force information. The muscle force prediction sub-model characterizes the correlation between active contraction force, passive contraction force, and muscle force of the muscle.

[0119] Figure 6 shows a mechanical properties diagram of a simulated musculoskeletal model according to an embodiment of the present invention, Figure 7 shows a model diagram of a tendon unit according to an embodiment of the present invention, Figure 8 The figure shows the activation dynamics and contraction dynamics composition diagram of the simulated musculoskeletal model according to an embodiment of the present invention.

[0120] It is additionally stated that Figure 6 a(t) in represents the muscle activation value, is the maximum static force that the muscle can achieve under ideal conditions (optimal length, full activation) when actively contracting; represents the characteristic strain of the tendon, embodying the property of the tendon in the transition from nonlinear to linear mechanical response during loading; is the characteristic strain of the passive elastic component of the muscle, embodying the property of the muscle in the transition from the passive elastic force to the significant increase in deformation. Figure 7 SE in the equation represents a series elastic element; PE represents a parallel elastic element; CE represents a contractile element; F mt (t) represents muscle force information; l t (t) represents tendon length; l mt (t) represents muscle unit length information; φ(t) represents the pennation angle; l m (t) represents instantaneous muscle fiber length information; t represents time. Figure 8 U stim in the equation represents the electrical stimulation intensity; f stim represents the electrical stimulation frequency; S u represents the first scaling factor; S f represents the second scaling factor; e(t) represents the neural activation value; a(t) represents the muscle activation value; l mt (t) represents muscle unit length information; V mt (t) represents muscle unit velocity information; F mt (t) represents muscle force information.

[0121] As shown in FIG. 1, the active contraction force prediction model is as follows: Figures 6 to 8

[0122] (15)

[0123] where F ce characterizes the target active contraction force information; L characterizes the target muscle fiber length information, which can be the instantaneous muscle fiber length information l m under the condition that the muscle activation value of the target muscle is a, i.e., the ratio of the instantaneous muscle fiber length information l m0 to the optimal muscle fiber length information l m can be calculated according to the tendon model; the optimal muscle fiber length information l m0 can be obtained by default according to the musculoskeletal model in combination with an annealing algorithm; f(L) characterizes the active contraction force information corresponding to the target muscle fiber length information, i.e., the force-length relationship of the active contraction component; v characterizes the target muscle fiber velocity information, which is defined as the ratio of the instantaneous muscle fiber contraction velocity v m to the maximum muscle fiber contraction velocity v max ; the maximum muscle fiber contraction velocity v max ​It can be the default value of the musculoskeletal model, or it can be optimized by combining the annealing algorithm; the instantaneous muscle fiber contraction velocity v m Obtained by differentiating the muscle fiber length; f(v) represents the active contraction force information corresponding to the target muscle fiber velocity information, that is, the force-velocity relationship of the active contraction component; F mx Indicates the predetermined maximum isovolumetric contraction force; a indicates the muscle activation value. In this embodiment of the present invention, the instantaneous muscle fiber contraction velocity v can be obtained by reversing the force-velocity function (corresponding to the above f(v)). m , and it can be used as the new value for the next integration step, and the instantaneous muscle fiber length and tendon length curves can be obtained by such a cycle.

[0124] The passive contraction force prediction model is as follows:

[0125] (16)

[0126] Among them, F pe Represents the target passive contraction force information; L represents the target muscle fiber length information; F mx Characterizes the predetermined maximum isovolumetric contraction force; f p (L) Represents the passive contraction force information corresponding to the target muscle fiber length information.

[0127] The muscle force prediction sub-model is shown below:

[0128] (17)

[0129] Among them, F mt represents muscle force information; φ represents the pennation angle, which can be obtained through experiments. The pennation angle represents the angle between the muscle fiber arrangement direction of the target muscle and the tendon axis.

[0130] Specifically, the muscle force prediction sub-model can also be expressed by the following formula:

[0131] (18)

[0132] Among them, F mt Represents muscle force information, that is, the force exerted by the muscle unit of the target muscle; F t Characterizes the force exerted by the tendon; F m represents the force generated by the muscle fiber; φ represents the pennation angle. For example, the pennation angle can be calculated using the following formula:

[0133] (19)

[0134] Where φ represents the feathering angle; l m Indicates instant muscle fiber length information; lm0 represents the optimal muscle fiber length information; φ0 represents the feathering angle when the optimal muscle fiber length information is obtained, which can be determined according to the default value of the muscle-bone model or obtained based on the annealing algorithm.

[0135] In addition, the present invention can also calculate muscle fiber length in other ways, such as through a simplified tendon model, a rigid tendon (ST) model that reduces calculation time, and an integral elastic tendon (IET) model.

[0136] Taking the integrated elastic tendon (IET) model as an example, the tendon force dynamics equations are first solved by numerically integrating a set of ordinary differential equations. The integrated elastic tendon (IET) model calculates the muscle fiber length by the forward integral of the muscle fiber velocity. The initial value of the fiber velocity is first estimated by an optimization procedure, which distributes the total tendon unit velocity between the fiber and the tendon. The instantaneous muscle fiber contraction velocity v is then calculated using the Runge-Kutta-Fehlberg algorithm. m Integrate and calculate the instantaneous muscle fiber length information l m and tendon length l t The calculation method of instant muscle fiber length information is as follows:

[0137] (20)

[0138] Among them, l m Indicates instant muscle fiber length information; l mt Indicates muscle unit length information; l t represents the tendon length; φ represents the pennation angle.

[0139] Then the tendon strain can be calculated, and the specific expression is as follows:

[0140] (twenty one)

[0141] Where, ε represents the strain of the tendon; L t Indicates tendon length; L ts The strain of the tendon can then be used to calculate the force exerted by the tendon. Then, the force F exerted by the tendon can be calculated using the force-strain relationship of the tendon. t, then combined with muscle activation to calculate the active and passive components of force. Because the active contractile component and the passive elastic component are in parallel and in series with the tendon elastic component, the relationship between their forces is as described in the muscle force prediction model described above. Based on this, inverting the muscle force-velocity function yields the instantaneous muscle fiber contraction velocity Vm, which is then used as the new value for the next integration step. This cycle of calculations yields the current muscle fiber length and tendon length curve.

[0142] A more robust implementation of the muscle-tendon (MTU) model is the balanced elastic-tendon (EET) model. This model uses the van Wengaarden-Dekker-Brent optimization procedure to find the roots of the aforementioned muscle force prediction model. The muscle force is then derived from the tendon force-strain relationship. By combining the aforementioned equations, tendon strain is expressed as a function of muscle fiber length. The muscle fiber velocity obtained during the muscle force calculation can be used as a numerical derivative of the muscle fiber length. This implementation provides a robust method for resolving fiber length. Unlike the integrated elastic-tendon (IET) model, it always maintains equilibrium in the muscle-tendon unit, resulting in a more reliable solution for fiber length.

[0143] In addition, the rigid tendon (ST) model is a simplified model used to estimate the muscle-tendon unit (MTU) force, in which the tendon can be regarded as an element with infinite stiffness and a tendon length l. t Fixed to the relaxed length l ts This simplified processing method reduces the variables and complexity in the calculation process, thereby effectively reducing the calculation time and improving the calculation efficiency.

[0144] At this point, the muscle fiber length information can be obtained based on the tendon model, and then the magnitude of the force generated by the active contraction component and passive elastic component of the muscle fiber can be solved based on the obtained muscle fiber length, and finally the muscle force generated by the muscle unit can be obtained.

[0145] Figure 9 A flow chart of a method for determining joint torque information according to an embodiment of the present invention is shown.

[0146] like Figure 9 As shown, the method of this embodiment may include operations S910 to S950.

[0147] In operation S910 , electrical stimulation parameters corresponding to the target object, a motion trajectory formed based on marker points, and a ground reaction force are collected.

[0148] In operation S920 , target muscle fiber length information and target muscle fiber velocity information of the target object are acquired.

[0149] In operation S930 , model parameters of the neural activation model, the muscle activation model, and the muscle force prediction model are calibrated.

[0150] In operation S940 , muscle force information is predicted using the neural activation model, the muscle activation model, and the muscle force prediction model.

[0151] In operation S950 , the muscle force information is processed using the joint torque model to obtain joint torque information.

[0152] At this point, the real-time muscle fiber length can be determined, and then the magnitude of the force generated by the active contraction component and the passive elastic component of the muscle fiber can be determined based on the obtained muscle fiber length, and then the muscle force generated by the muscle unit can be determined.

[0153] Figure 10 A block diagram of an electronic device suitable for implementing a muscle force prediction method under functional electrical stimulation according to an embodiment of the present invention is shown.

[0154] like Figure 10 As shown, an electronic device 1000 according to an embodiment of the present invention includes a processor 1001, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage unit 1008 into a random access memory (RAM) 1003. Processor 1001 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. Processor 1001 may also include onboard memory for caching purposes. Processor 1001 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present invention.

[0155] Various programs and data required for the operation of electronic device 1000 are stored in RAM 1003. Processor 1001, ROM 1002, and RAM 1003 are connected to each other via bus 1004. Processor 1001 performs various operations according to the method flow of an embodiment of the present invention by executing the programs in ROM 1002 and / or RAM 1003. It should be noted that the programs may also be stored in one or more memories other than ROM 1002 and RAM 1003. Processor 1001 may also perform various operations according to the method flow of an embodiment of the present invention by executing the programs stored in the one or more memories.

[0156] According to an embodiment of the present invention, electronic device 1000 may further include an input / output (I / O) interface 1005, which is also connected to bus 1004. Electronic device 1000 may also include one or more of the following components connected to I / O interface 1005: an input section 1006 including a keyboard, mouse, etc.; an output section 1007 including devices such as a cathode ray tube (CRT), liquid crystal display (LCD), and speakers; a storage section 1008 including a hard disk; and a communication section 1009 including a network interface card such as a LAN card or modem. Communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to I / O interface 1005 as needed. Removable media 1011, such as a magnetic disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed in drive 1010 as needed, so that computer programs read from the removable media can be installed into storage section 1008 as needed.

[0157] The present invention also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments, or may exist independently and not incorporated into the device / apparatus / system. The computer-readable storage medium carries one or more programs, which, when executed, implement the method according to the embodiments of the present invention.

[0158] According to an embodiment of the present invention, a computer-readable storage medium may be a non-volatile computer-readable storage medium, and may include, for example, but not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to an embodiment of the present invention, a computer-readable storage medium may include the ROM 902 and / or RAM 1003 described above, and / or one or more memories other than ROM 1002 and RAM 1003.

[0159] Embodiments of the present invention also include a computer program product comprising a computer program containing program code for executing the method shown in the flowchart. When the computer program product is executed in a computer system, the program code causes the computer system to implement the method for predicting muscle force under functional electrical stimulation provided by an embodiment of the present invention.

[0160] The above-described functions of the system / device defined in the system / apparatus of the embodiments of the present application are performed when the computer program is executed by the processor 1001. According to the embodiments of the present application, the system, apparatus, module, unit, etc. described above can be implemented by the computer program modules.

[0161] In one embodiment, the computer program can be stored in a tangible storage medium, such as an optical, magnetic, or other memory on a hard disk drive, solid-state drive, or other storage device. In another embodiment, the computer program can be transmitted over a network, using a radio wave, microwave, or other form of electromagnetic radiation, for example, and can be downloaded and installed by the communication portion 1009 and / or installed from the removable medium 1011. The program code contained in the computer program can be transmitted using any suitable network medium, including, but not limited to, wireless, wireline, optical, etc., or any suitable combination of the foregoing.

[0162] In such an embodiment, the computer program can be downloaded and installed from the network by the communication portion 1009 and / or installed from the removable medium 1011. When the computer program is executed by the processor 1001, the above-described functions of the system defined in the system of the embodiments of the present application are performed. According to the embodiments of the present application, the system, apparatus, module, unit, etc. described above can be implemented by the computer program modules.

[0163] According to the embodiments of the present application, the program code for carrying out the computer program provided by the embodiments of the present application can be written in any combination of one or more programming languages, and can specifically be implemented using a high-level procedural and / or object-oriented programming language, and / or an assembly / machine language. The programming language includes, but is not limited to, a programming language such as Java, C++, Python, "C" language, or a similar programming language. The program code can be executed entirely on the user computing device, partially on the user device, partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any kind of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect through the Internet).

[0164] The computer program product of the present application can be a computer program product comprising a computer readable storage medium and a computer program mechanism embedded in the computer readable storage medium. Such computer program product can further include a computer readable storage medium and program means for causing a processor or other programmable processing apparatus to function in a particular manner, such that the computer program mechanism embedded in the computer readable storage medium can be used to actually effect the apparatus functions.

[0165] Those skilled in the art will appreciate that the features recited in the various embodiments of the present application can be combined and / or integrated in a variety of ways, even if such combinations or integrations are not expressly noted in the present application. In particular, the features recited in the various embodiments of the present application can be combined and / or integrated in a variety of ways without departing from the spirit and scope of the present application. All such combinations and / or integrations are within the scope of the present application.

[0166] The embodiments of the present application have been described above. However, these embodiments are merely intended for illustration, and are not intended to limit the scope of the present application. Although the embodiments are described separately above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Various alternatives and modifications can be made to the embodiments of the present application by those skilled in the art without departing from the scope of the present application, and such alternatives and modifications are intended to fall within the scope of the present application.

Claims

1. A method for predicting muscle force under functional electrical stimulation, characterized in that: The muscle force prediction method comprises: Obtaining muscle fiber length information and muscle fiber velocity information of a target subject, wherein the muscle fiber length information and the muscle fiber velocity information respectively represent the muscle fiber length and muscle fiber velocity of the target muscle when the target muscle of the target subject is electrically stimulated according to electrical stimulation parameters and the target subject performs an action task; Processing the electrical stimulation parameters using a neural activation model to obtain a neural activation value; the neural activation model represents a correlation between the electrical stimulation parameters and the neural activation value; the neural activation value represents a degree of neural activation of the nerve by the electrical stimulation parameters; Processing the nerve activation value using a muscle activation model to obtain a muscle activation value; the muscle activation model represents a correlation between the nerve activation value and the muscle activation value; the muscle activation value represents the activation level of the target muscle when the nerve of the target muscle is at the nerve activation level; The muscle force prediction model is used to process the muscle fiber length information, the muscle fiber velocity information and the muscle activation value to predict the muscle force information of the target muscle.

2. The muscle force prediction method according to claim 1, wherein: The muscle fiber length information includes instant muscle fiber length information and optimal muscle fiber length information; The muscle fiber speed information includes instant muscle fiber speed information and optimal muscle fiber speed information; The method of using a muscle force prediction model to process the muscle fiber length information, the muscle fiber velocity information, and the muscle activation value to predict the muscle force information of the target muscle includes: Normalizing the instantaneous muscle fiber length information using the optimal muscle fiber length information to obtain normalized target muscle fiber length information; Normalizing the instantaneous muscle fiber velocity information using the optimal muscle fiber velocity information to obtain normalized target muscle fiber velocity information; The muscle force prediction model is used to process the target muscle fiber length information, the target muscle fiber velocity information, and the muscle activation value to obtain the muscle force information.

3. The muscle force prediction method according to claim 2, characterized in that: The muscle force prediction model includes an active contraction force prediction model, a passive contraction force prediction model and a muscle force prediction sub-model; The step of using the muscle force prediction model to process the target muscle fiber length information, the target muscle fiber velocity information, and the muscle activation value to obtain the muscle force information includes: The active contraction force prediction model is used to process the target muscle fiber length information, the target muscle fiber velocity information, and the muscle activation value to obtain target active contraction force information of the target muscle; the active contraction force prediction model characterizes the correlation between the predetermined muscle activation value, the muscle fiber length, the muscle fiber velocity, and the active contraction force of the actively contracting muscle unit in the muscle; The passive contraction force prediction model is used to process the target muscle fiber length information to obtain target passive contraction force information of the target muscle; the passive contraction force prediction model characterizes the correlation between the muscle fiber length and the passive contraction force of the passive contraction muscle unit in the muscle; The target active contraction force information and the target passive contraction force information are processed by the muscle force prediction sub-model to obtain the muscle force information; the muscle force prediction sub-model characterizes the correlation between the active contraction force, the passive contraction force and the muscle force.

4. The muscle force prediction method according to any one of claims 1 to 3, characterized in that: The electrical stimulation parameters include electrical stimulation intensity and electrical stimulation frequency; The step of processing the electrical stimulation parameters using a neural activation model to obtain a neural activation value includes: calculating a first proportional factor based on the electrical stimulation intensity and a predetermined threshold; A second proportional factor is calculated based on the electrical stimulation frequency, a first predetermined constant, a second predetermined constant, a predetermined frequency constant and a predetermined relationship parameter; the first predetermined constant is calculated based on the second predetermined constant, the predetermined frequency constant and the predetermined relationship parameter; the predetermined frequency constant is calculated based on the second predetermined constant, a predetermined critical fusion frequency and the predetermined relationship parameter; the predetermined critical fusion frequency represents a critical frequency for stimulating muscle contraction; the second predetermined constant is a ratio of a predetermined maximum contraction force of the target muscle to a contraction force of the target muscle generated by electrical stimulation when the electrical stimulation frequency is the predetermined critical fusion frequency; the predetermined relationship parameter is used to represent a relationship between a predetermined electrical stimulation frequency and muscle force; The first scale factor and the second scale factor are processed using the neural activation model to obtain the neural activation value.

5. The muscle force prediction method according to any one of claims 1 to 3, characterized in that: The model parameters of the neural activation model, the muscle activation model, and the muscle force prediction model are determined by the following method: In the i-th iteration round, based on the i-th random perturbation, the i-1-th model parameter is updated to obtain the candidate model parameter; the i-th random perturbation is determined based on the i-th parameter optimization step size and the i-th perturbation random value randomly generated in the i-th iteration round; Obtaining an i-th electrical stimulation parameter, and determining a candidate joint torque based on the i-th electrical stimulation parameter and the candidate model parameters; determining a first difference value of the candidate joint torque relative to the experimental joint torque; The experimental joint torque represents an actual value of the joint torque measured during the walking test of the target subject; When the first difference value is smaller than the second difference value, determining the candidate joint torque as the i-th joint torque, and determining the candidate model parameter as the i-th model parameter; The second difference value is a difference value of the (i-1)th joint torque relative to the experimental joint torque; According to a predetermined step size adjustment strategy, based on an i-th acceptance rate of the candidate joint torque and a predetermined acceptance rate threshold, updating the i-th parameter optimization step size to obtain an i+1-th parameter optimization step size; The i-th acceptance rate represents the proportion of the joint torque determined according to the candidate joint torque in the i-th iteration round to the i-th joint torque; updating the i-th annealing temperature parameter based on the update number of the i+1-th parameter optimization step to obtain the i+1-th annealing temperature parameter; When the i+1th annealing temperature parameter is less than or equal to a predetermined temperature threshold, or the difference between the first difference value and the second difference value is less than a predetermined difference threshold, or i+1=m, the i-th model parameter is determined as the model parameter of the neural activation model, the muscle activation model and the muscle force prediction model, where m is a predetermined round threshold.

6. The muscle force prediction method according to claim 5, characterized in that: Determining a candidate joint torque based on the i-th electrical stimulation parameter and the candidate model parameter includes: determining candidate muscle force information based on the i-th electrical stimulation parameter and the candidate model parameter; The candidate joint torque is obtained according to the candidate muscle force information.

7. The muscle force prediction method according to claim 5, characterized in that: Also includes: When the first difference value is greater than or equal to the second difference value and the acceptance probability of the candidate joint torque is greater than an i-th random number randomly generated in the i-th iteration round, determining the candidate joint torque as the i-th joint torque, and determining the candidate model parameter as the i-th model parameter; When the first difference value is greater than or equal to the second difference value and the acceptance probability of the candidate joint torque is less than or equal to the i-th random number, the i-1-th joint torque is determined as the i-th joint torque, and the i-1-th model parameter is determined as the i-th model parameter.

8. The muscle force prediction method according to any one of claims 1 to 3, characterized in that: The coefficients of the muscle activation model are determined based on a first predetermined time constant, a second predetermined time constant, and a third predetermined time constant; the first predetermined time constant represents a predetermined time constant when electrically stimulating a muscle; The second predetermined time constant represents a predetermined time constant of the time it takes for the muscle force to rise from a base value to a peak value after the muscle is activated; The third predetermined time constant represents a predetermined time constant of the time it takes for the muscle force to drop from a peak value to a base value after the muscle activation ends.

9. The muscle force prediction method according to any one of claims 1 to 3, characterized in that: The muscle fiber length information and muscle fiber velocity information are determined based on the muscle-tendon unit length information and the tendon model. The muscle-tendon unit length information is collected when the target muscle of the target object is electrically stimulated according to the electrical stimulation parameters and the target object performs the action task, and the motion trajectory information and the ground reaction force information are analyzed and determined based on the predetermined musculoskeletal model of the target object.

10. The muscle force prediction method according to any one of claims 1 to 3, characterized in that: There are multiple target muscles; the multiple target muscles are distributed around the target joint of the target object; The muscle force prediction method further comprises: Obtain muscle moment arm information for each target muscle; determining a joint torque of the target joint according to a predetermined weight of each target muscle, muscle force information of each target muscle, and muscle moment arm information of each target muscle; determining an error value of the joint torque relative to a predetermined joint torque; In a case where the error value does not satisfy a predetermined condition, the electrical stimulation parameter is adjusted based on the error value to obtain an adjusted electrical stimulation parameter.

Images