Exoskeleton joint torque control method and device, computer equipment, readable storage medium and program product

Abstract

The invention relates to an exoskeleton joint torque control method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring a joint angle, a joint angular velocity and a current joint torque of an exoskeleton joint, and acquiring an electromyographic signal of an exoskeleton wearer; obtaining an expected joint torque of the exoskeleton joint according to the joint angle, the joint angular velocity, the current joint torque and the electromyographic signal by using a joint torque prediction model; according to the expected joint torque and the current joint torque, the torque error of the exoskeleton joint is obtained; adjusting a control parameter of a proportional-integral-derivative controller according to the torque error to obtain an adjusted proportional-integral-derivative controller; and obtaining a control signal of the exoskeleton joint according to the torque error by utilizing an adjusted proportional integral derivative controller. By adopting the method, the control precision of the exoskeleton can be improved, and the high-precision operation requirements of converter stations and other application scenes can be met.

Description

Technical Field

[0001] The present application relates to the field of exoskeleton control technology, and in particular to an exoskeleton joint torque control method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Art

[0002] An exoskeleton robot is a device that can simulate, enhance or restore human movement ability. With the development of exoskeleton technology, exoskeleton robots are gradually being used in power facilities such as converter stations to reduce the workload of operators.

[0003] In the related art, a proportional-integral-differential controller is usually used to control the exoskeleton robot. However, due to the dynamic characteristics of the exoskeleton robot and the uncertainty of the application environment, the traditional exoskeleton control method often cannot meet the high-precision operation requirements of application scenarios such as converter stations. Summary of the invention

[0004] Based on this, it is necessary to provide an exoskeleton joint torque control method, device, computer equipment, computer readable storage medium and computer program product to address the above technical problems.

[0005] In a first aspect, the present application provides an exoskeleton joint torque control method, comprising:

[0006] Obtain the joint angle, joint angular velocity, current joint torque of the exoskeleton joints, and obtain the electromyographic signal of the exoskeleton wearer;

[0007] Using a joint torque prediction model, the expected joint torque of the exoskeleton joint is obtained according to the joint angle, the joint angular velocity, the current joint torque and the electromyographic signal;

[0008] Obtaining a torque error of the exoskeleton joint according to the expected joint torque and the current joint torque;

[0009] adjusting a control parameter of a proportional-integral-differential controller according to the torque error to obtain an adjusted proportional-integral-differential controller;

[0010] The adjusted proportional-integral-derivative controller is used to obtain a control signal of the exoskeleton joint according to the torque error.

[0011] In one embodiment, the control parameters of the proportional-integral-differential controller are adjusted according to the torque error to obtain an adjusted proportional-integral-differential controller, including: obtaining a first proportional coefficient, a first integral coefficient and a first differential coefficient of the proportional-integral-differential controller; adjusting the first proportional coefficient according to the torque error and the proportional adjustment parameter to obtain a second proportional coefficient; adjusting the first integral coefficient according to the error integral of the torque error and the integral adjustment parameter to obtain a second integral coefficient; adjusting the first differential coefficient according to the error differential of the torque error and the differential adjustment parameter to obtain a second differential coefficient; and obtaining the adjusted proportional-integral-differential controller according to the second proportional coefficient, the second integral coefficient and the second differential coefficient.

[0012] In one embodiment, before obtaining the adjusted proportional-integral-differential controller according to the second proportional coefficient, the second integral coefficient, and the second differential coefficient, the method includes: constructing a Lyapunov function of the torque error; using the Lyapunov function to verify the changing trend of the torque error according to the second proportional coefficient, the second integral coefficient, and the second differential coefficient; and when verifying that the changing trend of the torque error is a decrease in error, executing the step of obtaining the adjusted proportional-integral-differential controller according to the second proportional coefficient, the second integral coefficient, and the second differential coefficient.

[0013] In one embodiment, the method of obtaining the joint angle, joint angular velocity, and current joint torque of an exoskeleton joint and obtaining the electromyographic signal of an exoskeleton wearer includes: using an angle sensor and / or an encoder of the exoskeleton joint to obtain the joint angle of the exoskeleton joint; calculating the joint angular velocity of the exoskeleton joint according to the joint angle; measuring the current joint torque of the exoskeleton joint using a force sensor; and obtaining the electromyographic signal of the exoskeleton wearer using a surface electromyographic sensor.

[0014] In one embodiment, the joint torque prediction model is trained by the following steps: obtaining multiple groups of sample data of the exoskeleton joint and expected joint torque samples corresponding to each group of sample data; wherein each group of sample data includes joint angle samples, joint angular velocity samples, current joint torque samples and electromyographic signal samples of the exoskeleton wearer; constructing a network model including an input layer, a hidden layer and an output layer; and training the network model using a stochastic gradient descent method according to the multiple groups of sample data and the corresponding expected joint torque samples to obtain the joint torque prediction model.

[0015] In one embodiment, the network model is trained using the stochastic gradient descent method based on the multiple groups of sample data and the corresponding expected joint torque samples to obtain the joint torque prediction model, which includes: performing mean removal and standardization processing on each sample data in the multiple groups of sample data to obtain the processed multiple groups of sample data; based on the processed multiple groups of sample data, executing the step of training the network model using the stochastic gradient descent method based on the multiple groups of sample data and the corresponding expected joint torque samples to obtain the joint torque prediction model.

[0016] In a second aspect, the present application also provides an exoskeleton joint torque control device, comprising:

[0017] The first acquisition module is used to acquire the joint angle, joint angular velocity, current joint torque of the exoskeleton joints, and acquire the electromyographic signal of the exoskeleton wearer;

[0018] A torque prediction module, used for obtaining the expected joint torque of the exoskeleton joint according to the joint angle, the joint angular velocity, the current joint torque and the electromyographic signal by using a joint torque prediction model;

[0019] A second acquisition module, configured to obtain a torque error of the exoskeleton joint according to the expected joint torque and the current joint torque;

[0020] A parameter adjustment module, used for adjusting the control parameters of the proportional-integral-differential controller according to the torque error to obtain an adjusted proportional-integral-differential controller;

[0021] The third acquisition module is used to obtain the control signal of the exoskeleton joint according to the torque error by using the adjusted proportional integral derivative controller.

[0022] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0023] Obtain the joint angle, joint angular velocity, current joint torque of the exoskeleton joints, and obtain the electromyographic signal of the exoskeleton wearer;

[0024] Using a joint torque prediction model, the expected joint torque of the exoskeleton joint is obtained according to the joint angle, the joint angular velocity, the current joint torque and the electromyographic signal;

[0025] Obtaining a torque error of the exoskeleton joint according to the expected joint torque and the current joint torque;

[0026] adjusting a control parameter of a proportional-integral-differential controller according to the torque error to obtain an adjusted proportional-integral-differential controller;

[0027] The adjusted proportional-integral-derivative controller is used to obtain a control signal of the exoskeleton joint according to the torque error.

[0028] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0029] Obtain the joint angle, joint angular velocity, current joint torque of the exoskeleton joints, and obtain the electromyographic signal of the exoskeleton wearer;

[0030] Using a joint torque prediction model, the expected joint torque of the exoskeleton joint is obtained according to the joint angle, the joint angular velocity, the current joint torque and the electromyographic signal;

[0031] Obtaining a torque error of the exoskeleton joint according to the expected joint torque and the current joint torque;

[0032] adjusting a control parameter of a proportional-integral-differential controller according to the torque error to obtain an adjusted proportional-integral-differential controller;

[0033] The adjusted proportional-integral-derivative controller is used to obtain a control signal of the exoskeleton joint according to the torque error.

[0034] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0035] Obtain the joint angle, joint angular velocity, current joint torque of the exoskeleton joints, and obtain the electromyographic signal of the exoskeleton wearer;

[0036] Using a joint torque prediction model, the expected joint torque of the exoskeleton joint is obtained according to the joint angle, the joint angular velocity, the current joint torque and the electromyographic signal;

[0037] Obtaining a torque error of the exoskeleton joint according to the expected joint torque and the current joint torque;

[0038] adjusting a control parameter of a proportional-integral-differential controller according to the torque error to obtain an adjusted proportional-integral-differential controller;

[0039] The adjusted proportional-integral-derivative controller is used to obtain a control signal of the exoskeleton joint according to the torque error.

[0040] The above-mentioned exoskeleton joint torque control method, apparatus, computer equipment, computer-readable storage medium and computer program product first obtain the joint angle, joint angular velocity, current joint torque of the exoskeleton joint and the electromyographic signal of the exoskeleton wearer, and then use the joint torque prediction model to obtain the expected joint torque of the exoskeleton joint according to the joint angle, joint angular velocity, current joint torque and electromyographic signal, and then obtain the torque error of the exoskeleton joint according to the expected joint torque and the current joint torque, and then adjust the control parameters of the proportional integral differential controller according to the torque error to obtain an adjusted proportional integral differential controller, and use the adjusted proportional integral differential controller to obtain the control signal of the exoskeleton joint according to the torque error. This solution uses a deep learning model to accurately predict the expected joint torque in view of the special properties of the exoskeleton robot joint torque and the high-precision operation requirements of scenarios such as converter stations. Then, the proportional-integral-differential controller is dynamically adjusted according to the torque error between the expected joint torque and the current joint torque. The adjusted proportional-integral-differential controller can be used to perform real-time, accurate and stable control of the joint torque of the exoskeleton in a complex and uncertain environment, thereby improving work efficiency and safety, while reducing the workload of operating employees and meeting the high-precision operation requirements of the exoskeleton in application scenarios such as converter stations. Moreover, since this solution uses a proportional-integral-differential controller that is adaptively adjusted according to the torque error, it can avoid the problems of poor adaptability and complex operation caused by the traditional proportional-integral-differential controller due to the control parameters being fixed to preset values ​​or requiring manual adjustment. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0042] Figure 1 A schematic diagram of a flow chart of an exoskeleton joint torque control method in one embodiment;

[0043] Figure 2 A schematic diagram of a flow chart for adjusting a proportional-integral-derivative controller in one embodiment;

[0044] Figure 3 A schematic diagram of a flow chart of a training process of a joint torque prediction model in one embodiment;

[0045] Figure 4 A schematic flow chart of an exoskeleton joint torque control method in another embodiment;

[0046] Figure 5 is a structural block diagram of an exoskeleton joint torque control device in one embodiment;

[0047] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0049] In one embodiment, Figure 1 As shown, a method for controlling the torque of an exoskeleton joint is provided. This embodiment uses the method applied to a terminal as an example. It is understandable that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. The terminal can be an exoskeleton device. In this embodiment, the method includes the following steps:

[0050] Step S101, obtaining the joint angles, joint angular velocities, and current joint torques of the exoskeleton joints, and obtaining the electromyographic signals of the exoskeleton wearer.

[0051] The exoskeleton joint may be one of the joints of the exoskeleton device, such as an elbow joint. The joint angle may be the current angle of the exoskeleton joint, the joint angular velocity may be the current change rate of the joint angle of the exoskeleton joint, and the current joint torque may be the currently measured torque of the exoskeleton joint. The electromyographic signal of the exoskeleton wearer may be an electrical signal collected from the surface of the relevant muscle of the wearer corresponding to the exoskeleton joint.

[0052] Step S102, using the joint torque prediction model, the expected joint torque of the exoskeleton joint is obtained according to the joint angle, joint angular velocity, current joint torque and electromyographic signal.

[0053] The joint torque prediction model may be a trained neural network model, which can predict the expected joint torque based on the input joint angle, joint angular velocity, current joint torque and electromyographic signal. The expected joint torque may be the torque required to be applied to the exoskeleton joint in order to assist the exoskeleton wearer to achieve his or her movement intention.

[0054] In this step, the joint angle, joint angular velocity, current joint torque and electromyographic signal can be input into the joint torque prediction model to obtain the expected joint torque output by the model.

[0055] Step S103, obtaining the torque error of the exoskeleton joint according to the expected joint torque and the current joint torque.

[0056] Among them, according to the difference between the expected joint torque output by the model and the current joint torque measured, the torque error of the exoskeleton joint can be obtained. For example, the torque error can be expressed as , where represents the torque error, represents the current joint torque, represents the desired joint torque.

[0057] Step S104, adjusting control parameters of the proportional-integral-derivative controller according to the torque error to obtain an adjusted proportional-integral-derivative controller.

[0058] Among them, the Proportional Integration Differentiation (PID) controller can output the control signal of the exoskeleton joint according to the input torque error to control the movement of the exoskeleton joint to reduce the torque error. Among them, the control parameters of the proportional integral differential controller can include proportional parameters, integral parameters and differential parameters, and its control formula can be expressed as:

[0059]

[0060] In the formula, is the control signal output by the proportional-integral-derivative controller, is the scale parameter, is the torque error, is the integration parameter, is the differential parameter.

[0061] In this step, based on the torque error of the exoskeleton joint, the proportional parameter, integral parameter and differential parameter of the proportional-integral-differential controller can be adjusted respectively to make it more conducive to the continuous reduction of the torque error, thereby obtaining an adjusted proportional-integral-differential controller.

[0062] Step S105, using the adjusted proportional integral derivative controller to obtain a control signal of the exoskeleton joint according to the torque error.

[0063] By inputting the torque error into the adjusted proportional integral differential controller, a control signal output by the controller for the exoskeleton joint can be obtained. The control signal can be used to control the movement of the exoskeleton joint to reduce the torque error.

[0064] In the above-mentioned exoskeleton joint torque control method, in view of the special properties of the joint torque of the exoskeleton robot and the high-precision operation requirements of the converter station and other scenes, the deep learning model is used to accurately predict the expected joint torque, and then the proportional integral differential controller is dynamically adjusted according to the torque error between the expected joint torque and the current joint torque. Thus, the adjusted proportional integral differential controller can be used to perform real-time, accurate and stable control of the joint torque of the exoskeleton in a complex and uncertain environment, thereby improving the operating efficiency and safety of the converter station and reducing the workload of the operating staff. Moreover, since the proportional integral differential controller that is adaptively adjusted according to the torque error is used in this scheme, the problems of poor adaptability and complex operation caused by the traditional proportional integral differential controller due to the control parameters being fixed to preset values ​​or requiring manual adjustment can be avoided.

[0065] In an exemplary embodiment, Figure 2 As shown, the control parameters of the proportional-integral-differential controller are adjusted according to the torque error to obtain an adjusted proportional-integral-differential controller, including:

[0066] Step S201, obtaining a first proportional coefficient, a first integral coefficient and a first differential coefficient of a proportional-integral-derivative controller.

[0067] Specifically, when adjusting the proportional-integral-differential controller, the first proportional coefficient, the first integral coefficient and the first differential coefficient of the controller may be obtained first, wherein the first proportional coefficient, the first integral coefficient and the first differential coefficient may be current control parameters of the proportional-integral-differential controller.

[0068] Step S202: adjusting the first proportional coefficient according to the torque error and the proportional adjustment parameter to obtain a second proportional coefficient.

[0069] Step S203: adjusting the first integral coefficient according to the error integral of the torque error and the integral adjustment parameter to obtain a second integral coefficient.

[0070] Step S204: adjusting the first differential coefficient according to the error differential of the torque error and the differential adjustment parameter to obtain a second differential coefficient.

[0071] When adjusting the control parameters of the proportional-integral-differential controller, based on the principle that a larger feedback torque is needed to correct the increased error, the proportional coefficient, integral coefficient, and differential coefficient of the controller can be adjusted in steps S202 to S204 respectively by the following formulas:

[0072]

[0073]

[0074]

[0075] In the formula, is the torque error, is the error integral of the torque error, is the error differential of the torque error; is the first proportionality coefficient, is the proportional adjustment parameter, is the second proportionality coefficient; is the first integral coefficient, is the integral adjustment parameter, is the second integral coefficient; is the first differential coefficient, is the differential adjustment parameter, The proportional adjustment parameter, the integral adjustment parameter, and the differential adjustment parameter may be constants preset according to the characteristics of the exoskeleton device.

[0076] Step S205, obtaining an adjusted proportional-integral-differential controller according to the second proportional coefficient, the second integral coefficient, and the second differential coefficient.

[0077] According to the second proportional coefficient, the second integral coefficient, and the second differential coefficient obtained in the above process, an adjusted proportional-integral-differential controller can be obtained. For example, the control formula of the adjusted proportional-integral-differential controller can be expressed as:

[0078]

[0079] In the formula, is the control signal output by the proportional-integral-derivative controller, is the torque error, is the second scale parameter, is the second integration parameter, is the second differential parameter.

[0080] In this embodiment, by adjusting the proportional coefficient, integral coefficient, and differential coefficient of the PID controller according to the torque error, the integral and derivative of the torque error, respectively, it is possible to achieve fine-tuning of the PID controller, which is conducive to obtaining a more accurate control signal and improving the accuracy of the exoskeleton joint movement.

[0081] In an exemplary embodiment, before obtaining an adjusted proportional-integral-differential controller according to the second proportional coefficient, the second integral coefficient, and the second differential coefficient, the steps include: constructing a Lyapunov function of the torque error; using the Lyapunov function to verify the changing trend of the torque error according to the second proportional coefficient, the second integral coefficient, and the second differential coefficient; and when it is verified that the changing trend of the torque error is a decrease in the error, executing the step of obtaining an adjusted proportional-integral-differential controller according to the second proportional coefficient, the second integral coefficient, and the second differential coefficient.

[0082] Specifically, in order to ensure the stability of the adjustment process, the Lyapunov stability theory can be applied in this embodiment to verify the change trend of the torque error. According to the Lyapunov stability theory, a Lyapunov function corresponding to the torque error can be constructed. For example, the function can be as follows:

[0083]

[0084] In the formula, is the value of the Lyapunov function, is the torque error.

[0085] Among them, by ensuring that the value of the Lyapunov function decreases at any time, it can be ensured that the rate of change of the torque error at each moment is less than 0 (the error decreases), thereby ensuring the stability of the adjustment process.

[0086] Based on this, in this embodiment, it is possible to verify whether the Lyapunov function decreases over time based on the adjusted second proportional coefficient, second integral coefficient, and second differential coefficient. If so, it can be determined that the change trend of the torque error corresponding to these control parameters is a decrease in error.

[0087] Among them, when it is verified that the change trend of the torque error is a decrease in the error, an adjusted proportional-integral-differential controller can be constructed based on the adjusted second proportional coefficient, second integral coefficient, and second differential coefficient.

[0088] In this embodiment, by using the Lyapunov function to verify the changing trend of the torque error, it can be ensured that the adjusted second proportional coefficient, second integral coefficient, and second differential coefficient can make the changing trend of the torque error decrease, which is beneficial to ensuring the stability of the system.

[0089] In an exemplary embodiment, the joint angle, joint angular velocity, and current joint torque of an exoskeleton joint are obtained, and the electromyographic signal of the exoskeleton wearer is obtained, including: using angle sensors and / or encoders of the exoskeleton joints to obtain the joint angles of the exoskeleton joints; calculating the joint angular velocity of the exoskeleton joints according to the joint angles; using force sensors to measure the current joint torque of the exoskeleton joints; and using surface electromyographic sensors to obtain the electromyographic signal of the exoskeleton wearer.

[0090] In this embodiment, data of the exoskeleton and the exoskeleton wearer can be collected through a variety of devices.

[0091] The corresponding joint angle can be collected by using an angle sensor arranged at the exoskeleton joint, or the joint angle can be collected by using an encoder of the exoskeleton joint. Optionally, when the joint angles collected by the joint sensor and the encoder are obtained at the same time, the two data can be fused to obtain the final joint angle of the exoskeleton joint. Based on the joint angles collected at multiple time points, the change speed of the joint angle of the exoskeleton joint can be calculated, thereby obtaining the current joint angular velocity of the exoskeleton joint.

[0092] The force sensor disposed at the exoskeleton joint may be used to collect the current joint torque corresponding to the exoskeleton joint. Exemplarily, the force sensor may be a torque sensor.

[0093] The surface electromyography sensors disposed on the muscle surface of the exoskeleton wearer can be used to collect the electromyography signals of the exoskeleton wearer. For example, the surface electromyography sensors can be disposed on the surface of the relevant muscles of the wearer corresponding to the exoskeleton joints.

[0094] In an exemplary embodiment, Figure 3 As shown in Figure 1, the joint torque prediction model is trained through the following steps:

[0095] Step S301, obtaining multiple groups of sample data of exoskeleton joints and expected joint torque samples corresponding to each group of sample data.

[0096] In this step, multiple groups of sample data of the exoskeleton joints and the expected joint torque samples corresponding to each group of sample data may be obtained.

[0097] Each set of sample data may include joint angle samples, joint angular velocity samples, current joint torque samples of the exoskeleton joints and myoelectric signal samples of the exoskeleton wearer collected at the same time. By setting devices such as force sensors, angle sensors, encoders, etc. at the exoskeleton joints, and setting devices such as surface electromyography sensors on the muscle surface of the exoskeleton wearer, and then running the exoskeleton system and letting the exoskeleton wearer move in a predetermined or natural way, the outputs of various data acquisition devices can be recorded during the movement process to obtain multiple sets of sample data. Each set of sample data may include joint angle samples, joint angular velocity samples, current joint torque samples of the exoskeleton joints and myoelectric signal samples of the exoskeleton wearer collected at the same time.

[0098] The expected joint torque samples corresponding to each set of sample data may be set according to the exoskeleton wearer's movement intention at the corresponding collection time, and may correspond to the torque that needs to be applied to the exoskeleton joint in order to help the exoskeleton wearer achieve his movement intention.

[0099] Optionally, before collecting sample data, various data collection devices may be adjusted and calibrated to ensure that their performance is stable and that they can accurately read and output signals.

[0100] Step S302, constructing a network model including an input layer, a hidden layer, and an output layer.

[0101] Exemplarily, in this step, a neural network model including an input layer, multiple hidden layers, and an output layer can be constructed. Optionally, the model can be a five-layer neural network model including an input layer, three hidden layers, and an output layer. Optionally, the hidden layer of the model can be a fully connected layer, which can be composed of a linear transformation and a nonlinear activation function (for example, ReLU). Among them, the output of the i-th hidden layer in the model can be expressed as: , where σ is the activation function, W is the weight parameter, b is the bias parameter, is the output of the i-th hidden layer, is the output of the i-1th hidden layer. Optionally, when constructing the network model, various parameters in the model can be set to preset initial parameter values.

[0102] Step S303, according to the multiple groups of sample data and the corresponding expected joint torque samples, the network model is trained using the stochastic gradient descent method to obtain a joint torque prediction model.

[0103] Among them, by utilizing multiple groups of sample data and the expected joint torque samples corresponding to each group of sample data, the error between the expected joint torque output by the model according to the sample data and the corresponding expected joint torque samples can be minimized as the training goal, and the stochastic gradient descent method can be used to iteratively train the aforementioned constructed network model.

[0104] Optionally, the error between the expected joint torque output by the model based on the sample data and the corresponding expected joint torque sample can be a mean square error, which can be expressed as: in, is the mean square error, is the number of samples in one training, is the expected joint torque sample, is the sample data, is the expected joint torque output by the network model based on the sample data.

[0105] When the stochastic gradient descent method is used to train the model, the gradient of the error to each model parameter can be calculated by the back propagation algorithm, and then the gradient is used to update each model parameter. For example, taking parameters W and b as examples, the corresponding update process can be expressed as: , where is the learning rate, is the gradient of E with respect to parameter W, is the gradient of E with respect to parameter b.

[0106] When the model error converges or a predetermined number of iterations is reached, the iteration can be stopped and the trained network model can be used as a joint torque prediction model.

[0107] In this embodiment, by using the stochastic gradient descent algorithm to train the joint torque prediction model, efficient training of the model can be achieved, which is conducive to quickly obtaining a joint torque prediction model that can accurately predict the expected joint torque.

[0108] In an exemplary embodiment, based on multiple groups of sample data and corresponding expected joint torque samples, the network model is trained using the stochastic gradient descent method to obtain a joint torque prediction model, which may include: performing mean removal and standardization processing on each sample data in the multiple groups of sample data to obtain multiple groups of processed sample data; based on the processed multiple groups of sample data, executing the step of training the network model using the stochastic gradient descent method based on the multiple groups of sample data and the corresponding expected joint torque samples to obtain a joint torque prediction model.

[0109] Before using the sample data to train the network model, the sample data may be preprocessed and then applied to the model training.

[0110] Optionally, for each type of data in the multiple groups of sample data (e.g., joint angle samples, joint angular velocity samples, current joint torque samples, and electromyographic signal samples), mean removal processing can be performed on each type of data. The average value of the same type of data can be calculated and subtracted from each data point to eliminate the constant offset in the data. Exemplarily, the mean removal processing process can be shown as follows:

[0111]

[0112]

[0113] In the formula, is the original data, is the average value of the original data, The data are processed after mean removal.

[0114] Optionally, the Z-Score standardization method can be used to standardize the data after the mean removal process, so that the data can be at the same level and avoid some dimensions dominating the training process due to excessive values. Exemplarily, the Z-Score processing process can be shown as follows:

[0115]

[0116] In the formula, For the data after mean removal, and They are The mean and standard deviation of .

[0117] In this way, multiple groups of sample data that have been processed by mean removal and standardization can be obtained, which can be applied to subsequent model training.

[0118] In this embodiment, by performing mean removal and standardization processing on the sample data before model training, it is helpful to reduce sample errors and improve the accuracy of the model.

[0119] In one embodiment, Figure 4 As shown, a method for controlling the torque of an exoskeleton joint in a specific embodiment is provided, which specifically includes the following steps:

[0120] Step S401, obtaining multiple groups of sample data of exoskeleton joints and expected joint torque samples corresponding to each group of sample data; performing mean removal processing and standardization processing on each sample data in the multiple groups of sample data to obtain multiple groups of processed sample data.

[0121] Step S402, constructing a network model including an input layer, a hidden layer, and an output layer.

[0122] Step S403, according to the multiple groups of sample data and the corresponding expected joint torque samples, the network model is trained using the stochastic gradient descent method to obtain a joint torque prediction model.

[0123] Step S404, obtaining the joint angles, joint angular velocities, and current joint torques of the exoskeleton joints, and obtaining the electromyographic signals of the exoskeleton wearer.

[0124] Step S405, using the joint torque prediction model, the expected joint torque of the exoskeleton joint is obtained according to the joint angle, joint angular velocity, current joint torque and electromyographic signal.

[0125] Step S406: Obtain the torque error of the exoskeleton joint according to the expected joint torque and the current joint torque.

[0126] Step S407, adjusting the control parameters of the proportional-integral-derivative controller according to the torque error to obtain an adjusted proportional-integral-derivative controller.

[0127] Step S408, using the adjusted proportional integral derivative controller to obtain a control signal of the exoskeleton joint according to the torque error.

[0128] Among them, in this embodiment, after completing the construction and training of the joint torque prediction model according to steps S401 to S403, the model and the adaptive adjustment strategy of the proportional integral differential controller can be integrated into the exoskeleton system. Then, during the operation of the exoskeleton, real-time data can be continuously collected, and by repeatedly executing steps S404 to S408, these data are input into the deep learning model to predict the joint torque, and then the torque error between the predicted expected joint torque and the current joint torque is input into the adaptive control strategy, the control parameters of the proportional integral differential controller are adjusted, and then the adjusted proportional integral differential controller is used to control the exoskeleton in real time according to the torque error.

[0129] The beneficial effects brought by the above embodiments are as follows:

[0130] First, this embodiment combines deep learning prediction and PID control strategy to give full play to their respective advantages in the control of the exoskeleton joint torque in the converter station. The deep learning model accurately predicts the joint torque, and the PID control strategy dynamically adjusts the parameters according to the prediction and measured errors to achieve real-time and accurate control of the torque.

[0131] Secondly, in the traditional method, PID control parameters generally need to be preset or manually adjusted, which has poor adaptability and complicated operation. In this embodiment, by designing an adaptive control strategy for PID control parameters, the influence of error, error integral and error change rate is fully considered, and the PID control parameters can be automatically adjusted in real time, further improving the control accuracy.

[0132] Thirdly, the control method in this embodiment does not need to rely on a specific hardware platform, and can be implemented by using only a device with a certain computing power and necessary sensor hardware, which makes the implementation process simple and has strong versatility and broad application prospects.

[0133] At the same time, the method in this embodiment can perform real-time, accurate and stable control of the joint torque of the exoskeleton in a complex and uncertain environment, thereby improving work efficiency and safety, reducing the workload of operating personnel, and fully meeting the operational requirements of the green converter station.

[0134] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0135] Based on the same inventive concept, the embodiment of the present application also provides an exoskeleton joint torque control device for implementing the exoskeleton joint torque control method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more exoskeleton joint torque control device embodiments provided below can refer to the limitations of the exoskeleton joint torque control method above, and will not be repeated here.

[0136] In an exemplary embodiment, Figure 5 As shown, an exoskeleton joint torque control device 500 is provided, comprising:

[0137] The first acquisition module 501 is used to acquire the joint angle, joint angular velocity, and current joint torque of the exoskeleton joints, and acquire the electromyographic signal of the exoskeleton wearer;

[0138] A torque prediction module 502, for obtaining the expected joint torque of the exoskeleton joint according to the joint angle, the joint angular velocity, the current joint torque and the electromyographic signal by using a joint torque prediction model;

[0139] A second acquisition module 503 is used to obtain a torque error of the exoskeleton joint according to the expected joint torque and the current joint torque;

[0140] A parameter adjustment module 504, configured to adjust control parameters of a proportional-integral-differential controller according to the torque error to obtain an adjusted proportional-integral-differential controller;

[0141] The third acquisition module 505 is used to obtain the control signal of the exoskeleton joint according to the torque error by using the adjusted proportional integral derivative controller.

[0142] In an exemplary embodiment, the parameter adjustment module 504 is used to: obtain a first proportional coefficient, a first integral coefficient and a first differential coefficient of a proportional-integral-differential controller; adjust the first proportional coefficient according to the torque error and the proportional adjustment parameter to obtain a second proportional coefficient; adjust the first integral coefficient according to the error integral and the integral adjustment parameter of the torque error to obtain a second integral coefficient; adjust the first differential coefficient according to the error differential and the differential adjustment parameter of the torque error to obtain a second differential coefficient; obtain the adjusted proportional-integral-differential controller according to the second proportional coefficient, the second integral coefficient and the second differential coefficient.

[0143] In an exemplary embodiment, the parameter adjustment module 504 is also used to: construct a Lyapunov function of the torque error; use the Lyapunov function to verify the changing trend of the torque error according to the second proportional coefficient, the second integral coefficient, and the second differential coefficient; when it is verified that the changing trend of the torque error is a decrease in error, execute the step of obtaining the adjusted proportional-integral-differential controller according to the second proportional coefficient, the second integral coefficient, and the second differential coefficient.

[0144] In an exemplary embodiment, the first acquisition module 501 is used to: use an angle sensor and / or an encoder of the exoskeleton joint to collect the joint angle of the exoskeleton joint; calculate the joint angular velocity of the exoskeleton joint based on the joint angle; use a force sensor to measure the current joint torque of the exoskeleton joint; and use a surface electromyography sensor to collect the electromyography signal of the exoskeleton wearer.

[0145] In an exemplary embodiment, the joint torque prediction model is trained by the following steps: obtaining multiple groups of sample data of the exoskeleton joint and expected joint torque samples corresponding to each group of sample data; wherein each group of sample data includes joint angle samples, joint angular velocity samples, current joint torque samples and electromyographic signal samples of the exoskeleton wearer; constructing a network model including an input layer, a hidden layer and an output layer; and training the network model using a stochastic gradient descent method based on the multiple groups of sample data and the corresponding expected joint torque samples to obtain the joint torque prediction model.

[0146] In an exemplary embodiment, the method includes: performing mean removal and standardization processing on each sample data in the multiple groups of sample data to obtain the processed multiple groups of sample data; and executing the step of training the network model based on the multiple groups of sample data and the corresponding expected joint torque samples using the stochastic gradient descent method to obtain the joint torque prediction model based on the multiple groups of sample data and the corresponding expected joint torque samples.

[0147] Each module in the above exoskeleton joint torque control device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each module.

[0148] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 6As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be realized through WIFI, a mobile cellular network, near field communication (NFC) or other technologies. When the computer program is executed by the processor, a method for controlling the torque of an exoskeleton joint is realized. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0149] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0150] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0151] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0152] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0153] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0154] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0155] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0156] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. An exoskeleton joint torque control method, characterized in that: The method comprises: Obtain the joint angle, joint angular velocity, current joint torque of the exoskeleton joints, and obtain the electromyographic signal of the exoskeleton wearer; Using a joint torque prediction model, the expected joint torque of the exoskeleton joint is obtained according to the joint angle, the joint angular velocity, the current joint torque and the electromyographic signal; Obtaining a torque error of the exoskeleton joint according to the expected joint torque and the current joint torque; adjusting a control parameter of a proportional-integral-differential controller according to the torque error to obtain an adjusted proportional-integral-differential controller; The adjusted proportional-integral-derivative controller is used to obtain a control signal of the exoskeleton joint according to the torque error.

2. The method according to claim 1, characterized in that The step of adjusting the control parameters of the proportional-integral-differential controller according to the torque error to obtain an adjusted proportional-integral-differential controller comprises: Obtaining a first proportional coefficient, a first integral coefficient, and a first differential coefficient of a proportional-integral-derivative controller; Adjusting the first proportional coefficient according to the torque error and the proportional adjustment parameter to obtain a second proportional coefficient; Adjusting the first integral coefficient according to the error integral of the torque error and the integral adjustment parameter to obtain a second integral coefficient; Adjusting the first differential coefficient according to the error differential and the differential adjustment parameter of the torque error to obtain a second differential coefficient; The adjusted proportional-integral-derivative controller is obtained according to the second proportional coefficient, the second integral coefficient, and the second differential coefficient.

3. The method according to claim 2, characterized in that Before obtaining the adjusted proportional-integral-differential controller according to the second proportional coefficient, the second integral coefficient, and the second differential coefficient, the method includes: constructing a Lyapunov function of the torque error; Using the Lyapunov function, verifying a change trend of the torque error according to the second proportional coefficient, the second integral coefficient, and the second differential coefficient; When it is verified that the change trend of the torque error is a decrease in error, a step of obtaining the adjusted proportional-integral-differential controller according to the second proportional coefficient, the second integral coefficient, and the second differential coefficient is performed.

4. The method according to claim 1, characterized in that: The method of obtaining the joint angle, joint angular velocity, and current joint torque of the exoskeleton joints and obtaining the electromyographic signal of the exoskeleton wearer includes: Using an angle sensor and / or an encoder of the exoskeleton joint to collect the joint angle of the exoskeleton joint; Calculating the joint angular velocity of the exoskeleton joint according to the joint angle; Measuring the current joint torque of the exoskeleton joint using a force sensor; The surface electromyography sensor is used to collect the electromyographic signal of the exoskeleton wearer.

5. The method according to any one of claims 1 to 4, characterized in that: The joint torque prediction model is trained by the following steps: Acquire multiple groups of sample data of the exoskeleton joint and expected joint torque samples corresponding to each group of sample data; wherein each group of sample data includes joint angle samples, joint angular velocity samples, current joint torque samples of the exoskeleton joint and myoelectric signal samples of the exoskeleton wearer; Construct a network model including input layer, hidden layer and output layer; According to the multiple groups of sample data and the corresponding expected joint torque samples, the network model is trained using a stochastic gradient descent method to obtain the joint torque prediction model.

6. The method according to claim 5, characterized in that The method comprises: training the network model using a stochastic gradient descent method according to the plurality of groups of sample data and the corresponding expected joint torque samples to obtain the joint torque prediction model; Performing mean removal processing and standardization processing on each sample data in the multiple groups of sample data to obtain the multiple groups of processed sample data; According to the processed multiple groups of sample data, the step of training the network model using the stochastic gradient descent method based on the multiple groups of sample data and the corresponding expected joint torque samples is performed to obtain the joint torque prediction model.

7. An exoskeleton joint torque control device, characterized in that: The device comprises: The first acquisition module is used to acquire the joint angle, joint angular velocity, current joint torque of the exoskeleton joints, and acquire the electromyographic signal of the exoskeleton wearer; A torque prediction module, used for obtaining the expected joint torque of the exoskeleton joint according to the joint angle, the joint angular velocity, the current joint torque and the electromyographic signal by using a joint torque prediction model; A second acquisition module, configured to obtain a torque error of the exoskeleton joint according to the expected joint torque and the current joint torque; A parameter adjustment module, used for adjusting the control parameters of the proportional-integral-differential controller according to the torque error to obtain an adjusted proportional-integral-differential controller; The third acquisition module is used to obtain the control signal of the exoskeleton joint according to the torque error by using the adjusted proportional integral derivative controller.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images