An online control method, device, equipment and medium of an exoskeleton robot

By performing rule fusion and high-dimensional spatial projection of the TS fuzzy model on the exoskeleton robot, and solving the consequent parameters in combination with training data, the relationship between the controller output torque and joint angle is determined, which solves the problem of poor control performance of the exoskeleton robot and realizes accurate online gait control.

CN116394229BActive Publication Date: 2026-03-03CENT SOUTH UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202310549615.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-05-16
Publication Date
2026-03-03
Estimated Expiration
2043-05-16

AI Technical Summary

Technical Problem

Exoskeleton robots have poor control performance, making it difficult to achieve accurate model building and effective control in complex nonlinear systems.

Method used

By modeling the exoskeleton robot based on fuzzy rules, a TS fuzzy model is obtained. The fuzzy rules are then fused and projected into a high-dimensional space. The consequent parameters are solved using training data to determine the relationship expression between the controller output torque and the joint angle, thereby realizing the online control of the exoskeleton robot.

Benefits of technology

Accurate modeling of exoskeleton robot systems under a limited set of rules improves the feasibility and accuracy of fuzzy control, enables accurate online gait control of exoskeleton robot systems, and enhances control performance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116394229B_ABST
    Figure CN116394229B_ABST
Patent Text Reader

Abstract

The application is suitable for the field of robot technology, and provides an online control method, device, equipment and medium for an exoskeleton robot. The online control method comprises the following steps: modeling the exoskeleton robot based on fuzzy rules to obtain a TS fuzzy model; fusing the fuzzy rules of the TS fuzzy model, and projecting the fused TS fuzzy model to a high-dimensional space to obtain an exoskeleton robot control model; solving the consequent parameters of the exoskeleton robot control model according to training data; determining a relationship expression between the controller output torque and the joint angle according to the solved consequent parameters; calculating the controller output torque of the expected joint angle at time T by using the relationship expression, and controlling the controller of the exoskeleton robot according to the calculated controller output torque at time T. The online control method can effectively improve the control performance of the exoskeleton robot.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of robotics technology, and in particular to an online control method, device, equipment and medium for an exoskeleton robot. Background Technology

[0002] In recent years, exoskeleton robots have been widely researched and applied in military and medical fields, providing wearers with protective functions and enhancing their capabilities, such as extending, supplementing, replacing, or strengthening human bodily functions, limb movement abilities, and load-bearing capacity. Exoskeleton robots require specific control methods to achieve normal human gait and assist human movement. Commonly used control methods include proportional-integral-derivative (PID) control, fuzzy control, gravity-compensated closed-loop control, and adaptive control. These methods have been successfully applied in many cases. However, because exoskeleton robots are typically complex nonlinear systems with multiple joints and strong human-machine coupling, and are subject to complex external environmental influences, it is often difficult to establish accurate models for these processes, resulting in suboptimal control performance of exoskeleton robots. Summary of the Invention

[0003] This invention provides an online control method, device, equipment, and medium for exoskeleton robots, which can solve the problem of poor control performance of exoskeleton robots.

[0004] In a first aspect, embodiments of the present invention provide an online control method for an exoskeleton robot, the control method comprising:

[0005] The TS fuzzy model is obtained by modeling the exoskeleton robot based on fuzzy rules.

[0006] The fuzzy rules in the TS fuzzy model are fused, and the fused TS fuzzy model is projected into a high-dimensional space to obtain the exoskeleton robot control model.

[0007] Based on the training data of the exoskeleton robot, solve for the consequent parameters of the exoskeleton robot control model;

[0008] Based on the obtained parameters of the successor, determine the expression relating the controller output torque to the joint angle;

[0009] The controller output torque is calculated based on the relationship between the controller output torque and the joint angle at time T, and the controller of the exoskeleton robot is controlled at time T according to the calculated controller output torque.

[0010] Optionally, the fused TS fuzzy model is as follows:

[0011]

[0012] in, This indicates the joint angles of the exoskeleton robot. , This represents the total number of fuzzy rules in the fused TS fuzzy model, where U is a vector related to joint angles and controller output torque. Let represent the membership degree of U to the k-th fuzzy rule. Let k be the kth fuzzy rule of the fused TS fuzzy model.

[0013] Optionally, the exoskeleton robot control model is as follows:

[0014]

[0015] in, , and All of these represent consequent parameters of the exoskeleton robot control model. Represents the projection function. Let represent the total membership degree of U with respect to all fuzzy sets in the k-th fuzzy rule, and let x represent the total number of fuzzy sets in each fuzzy rule. It represents the membership degree of U to the q-th fuzzy set in the k-th fuzzy rule.

[0016] Optionally, the training data of the exoskeleton robot includes N training samples and the joint angle of the exoskeleton robot at time T corresponding to each training sample. The training samples include: the joint angle of the exoskeleton robot at the n times before time T, and the controller output torque of the exoskeleton robot at time T.

[0017] Based on the training data of the exoskeleton robot, the consequent parameters of the exoskeleton robot control model are solved, including:

[0018] Construct the objective function:

[0019]

[0020] in, Describe the objective function. and This represents the consequent parameters of the exoskeleton robot control model. , This represents the total number of fuzzy rules in the fused TS fuzzy model. This represents the i-th training sample out of N training samples. , Let represent the joint angle of the exoskeleton robot corresponding to the i-th training sample at time T. Let represent the joint angle of the exoskeleton robot in the i-th training sample at time T-1. Let represent the joint angle of the exoskeleton robot at time Tn in the i-th training sample. Let n represent the controller output torque of the exoskeleton robot at time T in the i-th training sample, where n is... The total number of joint angles, Indicates modeling error. Represents the regularization factor. express The membership degree of the k-th fuzzy rule, Represents the projection function;

[0021] The objective function is solved using the Lagrange multiplier method;

[0022] The kernel function is used to perform matrix transformation on the solution of the objective function, and the transformed matrix is ​​then solved to obtain the consequent parameters of the exoskeleton robot control model. , .

[0023] Optionally, based on the obtained consequent parameters, the relationship expression between the controller output torque and the joint angle is determined, including:

[0024] Based on the consequent parameters obtained after solving, the k-th fuzzy rule of the exoskeleton robot control model is obtained;

[0025] Based on the kth fuzzy rule of the exoskeleton robot control model, the relationship expression between the controller output torque and the joint angle is obtained.

[0026] Optional, the k-th fuzzy rule of the exoskeleton robot control model for:

[0027]

[0028] in, Represents the Lagrange multipliers. Here, N represents the kernel function, and N represents the total number of training samples. express The membership degree of the k-th fuzzy rule, This represents the consequent parameters of the exoskeleton robot control model.

[0029] Optionally, based on the k-th fuzzy rule of the exoskeleton robot control model, the relationship expression between the controller output torque and the joint angle is obtained, including:

[0030] The k-th fuzzy rule of the exoskeleton robot control model is obtained by performing a first-order Taylor expansion and taking the derivative. for:

[0031]

[0032] in, This represents the joint angle of the exoskeleton robot at time T-1 under the k-th fuzzy rule. express The membership degree of the k-th fuzzy rule, Let represent the Lagrange multiplier, and N represent the total number of training samples. The width of the Gaussian function. This indicates that the exoskeleton robot under the k-th fuzzy rule is... Output torque at any moment This indicates that the exoskeleton robot in the i-th training sample under the k-th fuzzy rule... Output torque at any moment , , This indicates that under the k-th fuzzy rule, the exoskeleton robot... Output torque at any moment This indicates that under the k-th fuzzy rule, the exoskeleton robot... Output torque at any given moment;

[0033] in, ,in, This represents the joint angle of the exoskeleton robot at time T-2 under the k-th fuzzy rule. This represents the joint angle of the exoskeleton robot in the i-th training sample under the k-th fuzzy rule at time T-2. This represents the joint angle of the exoskeleton robot at time T-3 under the k-th fuzzy rule. Let represent the joint angle of the exoskeleton robot in the i-th training sample under the k-th fuzzy rule at time T-3. This represents the joint angle of the exoskeleton robot at time T-4 under the k-th fuzzy rule. Let represent the joint angle of the exoskeleton robot in the i-th training sample under the k-th fuzzy rule at time T-4;

[0034] Calculating each fuzzy rule after differentiation, the controller output torque under each fuzzy rule is obtained as follows:

[0035]

[0036] By weighting the controller output torque under each fuzzy rule, the relationship between the controller output torque and the joint angle is obtained as follows:

[0037]

[0038] in, This represents a preset constant value. Let represent the expected joint angle of the exoskeleton robot at time T under the k-th fuzzy rule. This represents the controller output torque of the exoskeleton robot at time T. This represents the output torque under the k-th fuzzy rule. The coefficients used for weighted summation.

[0039] Secondly, embodiments of this application provide an online control device for an exoskeleton robot, comprising:

[0040] The modeling module models the exoskeleton robot based on fuzzy rules, resulting in a TS fuzzy model.

[0041] The rule fusion module fuses the fuzzy rules in the TS fuzzy model and projects the fused TS fuzzy model into a high-dimensional space to obtain the exoskeleton robot control model.

[0042] The consequent parameter solving module solves for the consequent parameters of the exoskeleton robot control model based on the training data of the exoskeleton robot.

[0043] The relational expression determination module determines the relational expression between the controller output torque and the joint angle based on the obtained consequent parameters.

[0044] The exoskeleton robot control module uses a relational expression to calculate the controller output torque for the desired joint angle at time T, and then controls the exoskeleton robot's controller based on the calculated controller output torque at time T.

[0045] Thirdly, embodiments of this application provide a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-described online control method for the exoskeleton robot.

[0046] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described online control method for an exoskeleton robot.

[0047] The above-mentioned solution in this application has the following beneficial effects:

[0048] In the embodiments of this application, a TS fuzzy model is obtained by modeling the exoskeleton robot based on fuzzy rules. Then, the fuzzy rules in the TS fuzzy model are fused, and the fused TS fuzzy model is projected into a high-dimensional space to obtain the exoskeleton robot control model. The consequent parameters in the exoskeleton robot control model are then solved using training data. Based on the solved consequent parameters, the relationship expression between the controller output torque and joint angles is determined. Finally, the desired joint angle is substituted into the relationship expression to obtain the controller output torque. Control commands are then input to the exoskeleton robot according to the obtained controller output torque to control the exoskeleton robot. By fusing the TS fuzzy model and projecting it into a high-dimensional space, the exoskeleton robot system can be accurately modeled with a limited number of rules, ensuring the feasibility and accuracy of fuzzy control. Simultaneously, the consequent parameters of the exoskeleton robot control model obtained by modeling are solved using training data, and the relationship expression between the controller output torque and joint angles is determined based on the solution results. This allows for accurate calculation of the controller output torque based on the desired joint angle, achieving accurate online gait control of the exoskeleton robot system, achieving good control performance, and effectively improving the control performance of the exoskeleton robot system.

[0049] Other beneficial effects of this application will be described in detail in the following detailed description section. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 A flowchart illustrating an online control method for an exoskeleton robot provided in an embodiment of this application;

[0052] Figure 2 A block diagram of an online control method for a data-driven exoskeleton robot provided in an embodiment of this application;

[0053] Figure 3 A block diagram illustrating the construction of an exoskeleton robot control model provided in an embodiment of this application;

[0054] Figure 4 This is a graph showing the output joint angles and actual desired joint angles of the exoskeleton robot control model during the training phase in one embodiment of this application.

[0055] Figure 5This is an error curve diagram between the output joint angle and the actual desired joint angle of the exoskeleton robot control model during the training phase in one embodiment of this application.

[0056] Figure 6 This is a graph showing the controller output torque calculated during the testing phase and the actual controller output torque in one embodiment of this application.

[0057] Figure 7 This is an error curve diagram between the controller output torque calculated during the testing phase and the actual controller output torque in one embodiment of this application.

[0058] Figure 8 This is a graph showing the joint angles and actual joint angles of the exoskeleton robot during the simulation testing phase in one embodiment of this application.

[0059] Figure 9 This is an error curve diagram between the joint angles and actual joint angles of the exoskeleton robot during the simulation testing phase in one embodiment of this application.

[0060] Figure 10 A schematic diagram of the structure of an online control device for an exoskeleton robot provided in an embodiment of this application;

[0061] Figure 11 This is a schematic diagram of the structure of a terminal device provided in an embodiment of this application. Detailed Implementation

[0062] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0063] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0064] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0065] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0066] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0067] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0068] To address the issue of poor control performance in existing exoskeleton robots, this application provides an online control method for exoskeleton robots. This method involves modeling the exoskeleton robot based on fuzzy rules to obtain a TS fuzzy model, then fusing the fuzzy rules in the TS fuzzy model, projecting the fused TS fuzzy model into a high-dimensional space to obtain an exoskeleton robot control model, solving for the consequent parameters in the exoskeleton robot control model using training data, determining the relationship expression between the controller output torque and the joint angle based on the solved consequent parameters, and finally substituting the desired joint angle into the relationship expression to obtain the controller output torque. Control commands are then input to the exoskeleton robot according to the obtained controller output torque to control the exoskeleton robot. By fusing rules into the TS fuzzy model and projecting it into a high-dimensional space, the exoskeleton robot system can be accurately modeled with a small number of rules, ensuring the feasibility and accuracy of fuzzy control. At the same time, the consequent parameters of the exoskeleton robot control model obtained by training data are solved, and the relationship expression between the controller output torque and the joint angle is determined based on the solution results. This allows the controller output torque to be accurately calculated based on the desired joint angle, realizing accurate online gait control of the exoskeleton robot system, achieving good control effect, and effectively improving the control performance of the exoskeleton robot system.

[0069] The online control method for the exoskeleton robot provided in this application will be described by way of example below.

[0070] like Figure 1 As shown, the online control method for the exoskeleton robot provided in this application includes the following steps:

[0071] Step 11: Model the exoskeleton robot based on fuzzy rules to obtain the TS fuzzy model.

[0072] In some embodiments of this application, a relationship model between exoskeleton joint angles and controller output can be established using the traditional TS-fuzzy method, thereby achieving the modeling of the exoskeleton robot and obtaining a TS fuzzy model. Specifically, the expression of this TS fuzzy model is as follows:

[0073]

[0074] in, This represents the joint angles of the exoskeleton robot (i.e., the joint angles of the exoskeleton robot at time T). This represents the total number of fuzzy rules. express For the The membership degree of a fuzzy rule. Specifically, a vector related to the joint angle and the controller output torque. , This represents the joint angles of the exoskeleton robot at time T-1. This represents the joint angle of the exoskeleton robot at time Tn, where n is... The total number of joint angles, This represents the controller output torque of the exoskeleton robot at time T. and Both are vectors. For example, if... for A vector of dimension, then for 3D vector and Multiply them to get a number.

[0075] It is worth mentioning that when constructing a TS fuzzy model using the TS fuzzy method, the antecedent part is divided into rules based on the linear relationship between input and output, while the consequent part consists of linear equations to form a linear combination of the rules, giving the global output of the nonlinear system good linear descriptive characteristics. That is, by using the TS fuzzy method to construct a TS fuzzy model in this application, the TS fuzzy model can achieve high fitting accuracy, thereby realizing precise control of the exoskeleton robot.

[0076] Step 12: Fuse the fuzzy rules in the TS fuzzy model and project the fused TS fuzzy model into a high-dimensional space to obtain the exoskeleton robot control model.

[0077] In some embodiments of this application, the fuzzy rules in the TS fuzzy model are fused to obtain the fused TS fuzzy model. This is to avoid the number of rules being too large. The rules are fused and reduced by fusion, and multiple fuzzy rules are merged into one to reduce the number of fuzzy rules. This enables accurate modeling of the exoskeleton robot system under the condition of a small number of rules.

[0078] Although the above steps can reduce the number of fuzzy rules, enabling accurate modeling of exoskeleton robot systems with a small number of rules, the TS fuzzy model after rule fusion is nonlinear, which is not conducive to model building. Therefore, projecting the fused TS fuzzy model into a high-dimensional space can make the TS fuzzy model appear linear, eliminating the drawback of nonlinear TS fuzzy rules being unfavorable for model building.

[0079] In some embodiments of this application, a spatial kernel function can be used to project the fused TS fuzzy model onto a high-dimensional space to obtain an exoskeleton robot control model.

[0080] Step 13: Based on the training data of the exoskeleton robot, solve for the consequent parameters of the exoskeleton robot control model.

[0081] Step 12 reduces the number of fuzzy rules and eliminates the disadvantage of nonlinearity in model building, but the resulting exoskeleton robot control model contains two currently unknown consequent parameters. and This causes problems in the modeling process. Therefore, this step utilizes the training data from the exoskeleton robot to solve for the consequent parameters, allowing the consequent parameters to... and Transform it into something known, and solve the modeling problem.

[0082] It should be noted that the above training data consists of multiple training samples, including joint angles and controller output torque data of the exoskeleton robot at multiple historical moments. In some embodiments of this application, for joint angles, common joint angle acquisition methods can be used to acquire joint angles, such as data acquired through inertial measurement units or sensors; for controller output torque, the controller output torque can be acquired through controller output torque data fed back from the exoskeleton.

[0083] Step 14: Based on the obtained parameters of the successor, determine the expression relating the controller output torque to the joint angle.

[0084] In some embodiments of this application, the kth fuzzy rule of the exoskeleton robot control model can be obtained by substituting the solved consequent parameters into the exoskeleton robot control model, and then the relationship expression between the controller output torque and the joint angle can be obtained by calculating the kth fuzzy rule of the exoskeleton robot control model.

[0085] Step 15: Calculate the controller output torque for the desired joint angle at time T using the relational expression, and control the exoskeleton robot's controller at time T based on the calculated controller output torque.

[0086] The time T mentioned above can be a future time, the expected joint angle is the expected joint angle of the exoskeleton robot at this future time, and the controller output torque is the controller output torque calculated by substituting the expected joint angle into the relational expression.

[0087] It should be noted that by substituting the desired joint angle into the relational expression, the controller output torque can be calculated. Based on this, the calculated controller output torque can be used as the output torque of the exoskeleton robot's motors, thus controlling the exoskeleton robot to achieve the desired joint angle posture.

[0088] It is worth mentioning that by performing rule fusion on the TS fuzzy model and projecting it into a high-dimensional space, it is possible to accurately model the exoskeleton robot system under a small number of rules, ensuring the feasibility and accuracy of fuzzy control. At the same time, the consequent parameters of the exoskeleton robot control model obtained by training data are solved, and the relationship expression between the controller output torque and the joint angle is determined based on the solution results. This allows the controller output torque to be accurately calculated based on the desired joint angle, realizing accurate online gait control of the exoskeleton robot system, achieving good control effect, and effectively improving the control performance of the exoskeleton robot system.

[0089] The specific steps of step 12 described above will be illustrated below with reference to specific embodiments.

[0090] In some embodiments of this application, the specific implementation process of step 12 above includes the following steps:

[0091] Step 12.1: Perform rule fusion on the fuzzy rules in the TS fuzzy model, merging multiple fuzzy rules into one to reduce the number of fuzzy rules.

[0092] In some embodiments of this application, the original TS fuzzy model has H fuzzy rules, from 1 to... The fuzzy rules are merged into a new fuzzy rule. arrive One fuzzy rule is merged into a second new fuzzy rule, and so on, until... The H fuzzy rules are merged into a new k-th fuzzy rule. Specifically, the rule fusion calculation process is as follows:

[0093]

[0094] The fused TS fuzzy model is as follows:

[0095]

[0096] in, For the k-th fuzzy rule after fusion, Here, H represents the joint angles of the exoskeleton robot, H represents the total number of fuzzy rules before fusion, and R represents the total number of fuzzy rules in the fused fuzzy model. Let represent the membership degree of U with respect to the first fuzzy set. This represents the membership degree of U to the second fuzzy set. Let represent the membership degree of U to the R-th fuzzy set. Let represent the membership degree of U to the k-th fuzzy set. This represents a vector related to the joint angle and the controller output torque. express For the The membership degree of a fuzzy rule. and Both are vectors.

[0097] It is worth mentioning that merging multiple fuzzy rules into one to reduce the number of fuzzy rules can avoid an excessive number of rules and ensure accurate modeling of the exoskeleton robot system with a small number of rules.

[0098] Step 12.2: Project the fused TS fuzzy model onto a high-dimensional space to obtain the exoskeleton robot control model.

[0099] Fusion rules The nonlinear nature of the fuzzy rules hinders model building. To obtain linear fuzzy rules, a spatial kernel function can be used to... Projecting onto a high-dimensional space yields linear fuzzy rules. Simultaneously, by substituting the linear fuzzy rules into the fused TS fuzzy model, the exoskeleton robot control model is obtained:

[0100]

[0101] in, , and All represent consequent parameters of the exoskeleton robot control model, where U is a vector related to joint angles and controller output torque. Represents the projection function. Let represent the total membership degree of U with respect to all fuzzy sets in the k-th fuzzy rule, and q represent the dimension of the fuzzy variable under each fuzzy rule. It represents the membership degree of U to the q-th fuzzy set in the k-th fuzzy rule.

[0102] It is worth mentioning that by using a spatial kernel function to project the fused TS fuzzy model into a high-dimensional space, the TS fuzzy model can be made to appear linear, thus eliminating the disadvantage of nonlinear TS fuzzy models being unfavorable for model building.

[0103] The specific steps of step 13 described above will be illustrated below with reference to specific embodiments.

[0104] In some embodiments of this application, the training data of the exoskeleton robot includes N training samples and the joint angle of the exoskeleton robot at time T corresponding to each training sample. Each training sample includes: the joint angle of the exoskeleton robot at n times before time T, and the controller output torque of the exoskeleton robot at time T.

[0105] Based on the training data mentioned above, the specific implementation process of step 13 includes the following steps:

[0106] The first step is to solve for the consequent parameters of the exoskeleton robot control model. and Construct the objective function:

[0107]

[0108] in, Describe the objective function. , This represents the total number of fuzzy rules in the fused fuzzy model, where N represents the total number of training samples. This represents the i-th training sample out of N training samples. , Let represent the joint angle of the exoskeleton robot corresponding to the i-th training sample at time T. Let represent the joint angle of the exoskeleton robot in the i-th training sample at time T-1. Let n represent the joint angle of the exoskeleton robot at time Tn in the i-th training sample, where n is a constant. The total number of joint angles, Let represent the controller output torque of the exoskeleton robot at time T in the i-th training sample. Indicates modeling error. This represents the regularization factor, which represents the trade-off between approximate accuracy and generalization. express The membership degree of the k-th fuzzy rule, This represents the projection function.

[0109] The second step is to construct the Lagrange function.

[0110] The expression for the Lagrange function is as follows:

[0111]

[0112] in, It represents the Lagrange multiplier.

[0113] The solution to the objective function obtained using the Lagrange multiplier method is as follows:

[0114]

[0115] Perform matrix transformations on the solution of the objective function using kernel functions:

[0116]

[0117] Using the training data, the consequent parameters of the exoskeleton control model can be calculated using the above equation. and and Lagrange multipliers .

[0118] in, Represents an R-row, R-column zero matrix. This represents the membership function matrix obtained from the training data, where R represents the number of rules and N represents the number of training samples. R represents the transpose of the membership function matrix obtained from the training data, where R represents the number of rules and N represents the number of training samples. This represents the intermediate value obtained. The resulting N x N matrix This represents an R-row, 1-column matrix consisting of the consequent parameter b in each fuzzy rule. Represents Lagrange multipliers The resulting N x 1 matrix Represents an R-row, 1-column zero matrix. This represents an N x 1 matrix composed of joint angle data at time T from the collected training data. This represents an intermediate value obtained from known data. This represents the regularization factor, which represents the trade-off between approximate accuracy and generalization. Represents an N x N identity matrix. This represents the j-th training sample. The membership degree of the k-th fuzzy rule, Represents the kernel function.

[0119] In some embodiments of this application, the kernel function is defined as a radial basis function. , .

[0120] After the above steps, the consequent parameters are solved, allowing the consequent parameters to... and Transform into known parameters.

[0121] It is worth mentioning that the consequent parameters and By transforming the parameters into known parameters, the modeling problems caused by the presence of unknown parameters are resolved.

[0122] The specific steps of step 14 described above will be illustrated below with reference to specific embodiments.

[0123] In some embodiments of this application, step 14 is specifically implemented as follows:

[0124] Step 14.1: Based on the consequent parameters obtained after solving, the kth fuzzy rule of the exoskeleton robot control model is obtained, and the exoskeleton robot control model can be updated.

[0125] The solution obtained Substitute the fused linear fuzzy rules In the process, the k-th fuzzy rule of the exoskeleton robot control model is obtained as follows:

[0126]

[0127] Introducing radial basis functions Transform the above expression into:

[0128]

[0129] Then, after weighting by membership functions, the control model of the exoskeleton robot is obtained as follows:

[0130] .

[0131] Step 14.2: Based on the kth fuzzy rule of the exoskeleton robot control model, obtain the expression for the relationship between the controller output torque and the joint angle.

[0132] A first-order Taylor expansion is performed on the k-th fuzzy rule of the exoskeleton robot control model:

[0133]

[0134] Introducing radial basis functions Taking the first derivative of the intermediate term in the above equation, we get:

[0135]

[0136] in , .

[0137] in, This represents the joint angle of the exoskeleton robot at time T-2 under the k-th fuzzy rule. This represents the joint angle of the exoskeleton robot in the i-th training sample under the k-th fuzzy rule at time T-2. This represents the joint angle of the exoskeleton robot at time T-3 under the k-th fuzzy rule. Let represent the joint angle of the exoskeleton robot in the i-th training sample under the k-th fuzzy rule at time T-3. This represents the joint angle of the exoskeleton robot at time T-4 under the k-th fuzzy rule. This represents the joint angle of the exoskeleton robot in the i-th training sample under the k-th fuzzy rule at time T-4. This indicates that the exoskeleton robot under the k-th fuzzy rule is... The joint angle at any given moment.

[0138] Therefore, the k-th fuzzy rule of the exoskeleton robot control model can be transformed into:

[0139]

[0140] in, This represents the joint angle of the exoskeleton robot at time T under the k-th fuzzy rule. This represents the total membership degree of all fuzzy rules in the i-th training sample. Let represent the Lagrange multiplier, and N represent the total number of training samples. The width of the Gaussian function. This indicates that the exoskeleton robot under the k-th fuzzy rule is... Output torque at any moment This indicates that in the i-th training sample, under the k-th fuzzy rule, the exoskeleton robot... Output torque at any given moment.

[0141] , , This indicates that under the k-th fuzzy rule, the exoskeleton robot... Output torque at any moment This indicates that under the k-th fuzzy rule, the exoskeleton robot... Output torque at any given moment.

[0142] in, These are higher-order infinitesimal terms that satisfy the following relationship:

[0143]

[0144] It should be noted that during the control of exoskeleton robots, processes such as the exoskeleton's controller output and inertia cannot change too rapidly within a very small time interval. Therefore, higher-order infinitesimal terms can be reasonably ignored, and the above equation can be approximated as:

[0145]

[0146] Calculating each fuzzy rule after differentiation, the controller output torque under each fuzzy rule is obtained as follows:

[0147]

[0148] Finally, the controller output torque under each fuzzy rule is weighted to obtain the final expression for the relationship between the controller output torque and the joint angle:

[0149]

[0150] in, This represents a preset constant value. Let represent the expected joint angle of the exoskeleton robot at time T under the k-th fuzzy rule. This represents the controller output torque of the exoskeleton robot at time T. This represents the output torque under the k-th fuzzy rule. The coefficients used for weighted summation.

[0151] In some embodiments of this application, It is a very small number, obtained through real-time data calculation in actual control. To prevent the denominator from being 0, when it approaches 0 in a negative direction, set... For a very small negative number, when its positive value approaches 0, set It is a very small positive number.

[0152] It is worth mentioning that by determining the relational expression based on the consequent parameters, in subsequent control, given a desired joint angle, the relevant historical joint angles and historical controller outputs can also be known. Thus, the desired controller output can be calculated through the relational expression. This method takes into account both historical and real-time data, enabling accurate online gait control of the exoskeleton robot system, achieving good control results, and effectively improving the control performance of the exoskeleton robot system.

[0153] The online control method for the exoskeleton robot described above will be illustrated below with a specific example.

[0154] like Figure 2 As shown, a TS fuzzy model is obtained by modeling the exoskeleton robot system. This TS fuzzy model is obtained through rule fusion and by utilizing training data. , After solving for the consequent parameters using the spatial kernel function, the control model of the exoskeleton robot is obtained. The controller output torque under each fuzzy rule of the exoskeleton robot control model A Taylor expansion is performed to obtain a linear TS fuzzy model. Then, an inverse transform is performed on the linear TS fuzzy model to obtain the relationship expression between the controller output torque and the joint angle. This relationship expression is then input into the TS inverse controller. The input of the TS inverse controller is the joint angle. and The output terminal is the controller output torque. The controller outputs torque into the exoskeleton robot system to control the exoskeleton robot, causing it to output the desired joint angles. .

[0155] The construction process of the above exoskeleton robot control model is as follows: Figure 3As shown in the figure, a TS fuzzy model is obtained by modeling the exoskeleton robot system. Rule fusion is then performed on the TS fuzzy model. Indicate (Corresponding to U in the above text) for the first The membership degree of a fuzzy rule, and the above... Correspondingly, after fusion Become , express The membership degree of the k-th fuzzy rule. The fused TS fuzzy model is projected using a spatial kernel function, and the consequent parameters are solved using the training data to obtain the exoskeleton robot control model. The training data is the data input into the building block in the figure. and And the data shown in the angle model in the figure. .

[0156] In some embodiments of this application, the accuracy of the online control method provided in this application is verified by simulation: a set of collected data is used as training data to construct a TS fuzzy model of the exoskeleton robot, and after solving the parameters, an exoskeleton robot control model is obtained. The output joint angle of the exoskeleton robot control model is similar to the actual expected joint angle. Figure 4 As shown, the error between the output joint angle and the actual desired joint angle is as follows: Figure 5 As shown, the output joint angle curve of the exoskeleton robot control model almost coincides with the actual given desired joint angle curve. Based on this, another set of data containing the same type was used to test the online control method provided in this application. The controller output torque calculated using the aforementioned relationship expression between the controller output torque and joint angle, and the actual controller output torque are shown below. Figure 6 As shown, the error between the calculated controller output torque and the actual controller output torque is as follows: Figure 7 As shown, it can be seen that the controller output torque u(T) obtained using this relational expression is consistent with the actual controller output torque. The calculated controller output torque is then added to the model for simulation, and the resulting joint angles of the exoskeleton robot are shown below. Figure 8 As shown, the error between the joint angles of the exoskeleton robot and the actual joint angles is as follows: Figure 9 As shown, the joint angle curves of the exoskeleton robot almost coincide with the actual joint angle curves.

[0157] It should be noted that, Figure 4 In the diagram, "Training output" represents the output joint angles of the exoskeleton robot control model, and "Proposed method" represents the actual desired joint angles. Figure 5The "Training error" indicates the error between the output joint angle and the actual expected joint angle. Figure 6 In the figure, "Testing output" represents the calculated controller output torque, and "Proposed method" represents the actual controller output torque. Figure 7 The "Testing error" indicates the error between the calculated controller output torque and the actual controller output torque. Figure 8 The "Testing output" indicates the joint angles output by controlling the exoskeleton robot based on the obtained controller output torque, while the "Proposed method" indicates the actual joint angles. Figure 9 The "Testing error" in the text represents the error between the joint angles output by the exoskeleton robot and the actual joint angles. Figure 4 , 8 In the diagram, the x-axis represents the training and testing data points, and the y-axis represents the joint angle data value at time T calculated by the fuzzy model. Figure 5 , 9 In the diagram, the x-axis represents the training and testing data points, and the y-axis represents the error between the joint angle data value calculated by the fuzzy model at time T and the actual joint angle data value. Figure 6 In the diagram, the x-axis represents the test data points, and the y-axis represents the controller output value at time T calculated by the fuzzy model. Figure 7 In the diagram, the x-axis represents the test data points, and the y-axis represents the error between the controller output value calculated by the fuzzy model at time T and the actual controller output value.

[0158] Therefore, it can be seen that the online control method for the exoskeleton robot provided in this application embodiment can effectively control the exoskeleton robot to reach the expected position, effectively track the gait of the exoskeleton robot, and has good control performance.

[0159] The online control device for the exoskeleton robot provided in this application is described below as an example.

[0160] like Figure 10 As shown, this application embodiment provides an online control device for an exoskeleton robot. The online control device 1000 for the exoskeleton robot includes:

[0161] Modeling module 1001 models the exoskeleton robot based on fuzzy rules to obtain the TS fuzzy model;

[0162] The rule fusion module 1002 fuses the fuzzy rules in the TS fuzzy model and projects the fused TS fuzzy model into a high-dimensional space to obtain the exoskeleton robot control model.

[0163] The consequent parameter solving module 1003 solves the consequent parameters of the exoskeleton robot control model based on the training data of the exoskeleton robot.

[0164] The relation expression determination module 1004 determines the relation expression between the controller output torque and the joint angle based on the obtained consequent parameters.

[0165] The exoskeleton robot control module 1005 uses a relational expression to calculate the controller output torque for the desired joint angle at time T, and controls the controller of the exoskeleton robot at time T based on the calculated controller output torque.

[0166] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0167] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0168] like Figure 11 As shown, an embodiment of this application provides a terminal device, wherein the terminal device D10 of this embodiment includes: at least one processor D100 ( Figure 11 The diagram shows only one processor, a memory D101, and a computer program D102 stored in the memory D101 and executable on the at least one processor D100, wherein the processor D100 executes the computer program D102 to implement the steps in any of the above method embodiments.

[0169] Specifically, when the processor D100 executes the computer program D102, it models the exoskeleton robot based on fuzzy rules to obtain a TS fuzzy model. It then fuses the fuzzy rules in the TS fuzzy model, solves for the consequent parameters in the exoskeleton robot control model using training data, and determines the relationship between the controller output torque and joint angles. Finally, it substitutes the desired joint angles into the relationship expression to obtain the controller output torque, and inputs control commands to the exoskeleton robot according to the obtained controller output torque, thus controlling the exoskeleton robot. By fusing the TS fuzzy model and projecting it into a high-dimensional space, it is possible to accurately model the exoskeleton robot system with a limited number of rules, ensuring the feasibility and accuracy of fuzzy control. Simultaneously, by solving for the consequent parameters of the exoskeleton robot control model using training data and determining the relationship between the controller output torque and joint angles based on the solution results, it is possible to accurately calculate the controller output torque based on the desired joint angles, achieving accurate online gait control of the exoskeleton robot system, achieving good control performance, and effectively improving the control performance of the exoskeleton robot system.

[0170] The processor D100 can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0171] In some embodiments, the memory D101 may be an internal storage unit of the terminal device D10, such as a hard disk or memory of the terminal device D10. In other embodiments, the memory D101 may be an external storage device of the terminal device D10, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the terminal device D10. Furthermore, the memory D101 may include both internal and external storage units of the terminal device D10. The memory D101 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory D101 can also be used to temporarily store data that has been output or will be output.

[0172] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the various method embodiments above.

[0173] This application provides a computer program product that, when run on a terminal device, enables the terminal device to implement the steps described in the various method embodiments above.

[0174] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to the online control method device / terminal device of the exoskeleton robot, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0175] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0176] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0177] The above description is the preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this invention, and these improvements and modifications should also be considered within the scope of protection of this invention.

Claims

1. An online control method for an exoskeleton robot, characterized in that, include: The TS fuzzy model is obtained by modeling the exoskeleton robot based on fuzzy rules. The fuzzy rules in the TS fuzzy model are fused, and the fused TS fuzzy model is projected into a high-dimensional space to obtain the exoskeleton robot control model. Based on the training data of the exoskeleton robot, the consequent parameters of the exoskeleton robot control model are solved; Based on the obtained parameters of the successor, determine the expression relating the controller output torque to the joint angle; The controller output torque is calculated based on the expected joint angle at time T using the aforementioned relational expression, and the controller of the exoskeleton robot is controlled at time T based on the calculated controller output torque.

2. The online control method according to claim 1, characterized in that, The fused TS fuzzy model is as follows: ; in, This indicates the joint angles of the exoskeleton robot. , This represents the total number of fuzzy rules in the fused TS fuzzy model, where U is a vector related to joint angles and controller output torque. Let represent the membership degree of U to the k-th fuzzy rule. Let k be the kth fuzzy rule of the fused TS fuzzy model.

3. The online control method according to claim 2, characterized in that, The control model for the exoskeleton robot is as follows: ; in, , and All of these represent consequent parameters of the exoskeleton robot control model. Represents the projection function. Let represent the total membership degree of U with respect to all fuzzy sets in the k-th fuzzy rule, and let x represent the total number of fuzzy sets in each fuzzy rule. It represents the membership degree of U to the q-th fuzzy set in the k-th fuzzy rule.

4. The online control method according to claim 3, characterized in that, The training data of the exoskeleton robot includes N training samples and the joint angle of the exoskeleton robot at time T corresponding to each training sample. The training samples include: the joint angle of the exoskeleton robot at n times before time T, and the controller output torque of the exoskeleton robot at time T. The step of solving for the consequent parameters of the exoskeleton robot control model based on the training data of the exoskeleton robot includes: Construct the objective function: ; in, Denotes the objective function, and This represents the consequent parameters of the exoskeleton robot control model. , This represents the total number of fuzzy rules in the fused TS fuzzy model. This represents the i-th training sample out of N training samples. , Let represent the joint angle of the exoskeleton robot corresponding to the i-th training sample at time T. Let represent the joint angle of the exoskeleton robot in the i-th training sample at time T-1. Let represent the joint angle of the exoskeleton robot in the i-th training sample at time Tn. Let n represent the controller output torque of the exoskeleton robot in the i-th training sample at time T, where n is... The total number of joint angles, Indicates modeling error. Represents the regularization factor. express The membership degree of the k-th fuzzy rule, Represents the projection function; The objective function is solved using the Lagrange multiplier method; The solution to the objective function is transformed into a matrix using a kernel function, and the transformed matrix is ​​then solved to obtain the consequent parameters of the exoskeleton robot control model. , .

5. The online control method according to claim 4, characterized in that, The step of determining the relationship expression between the controller output torque and the joint angle based on the obtained consequent parameters includes: Based on the consequent parameters obtained after solving, the kth fuzzy rule of the exoskeleton robot control model is obtained; Based on the kth fuzzy rule of the exoskeleton robot control model, the relationship expression between the controller output torque and the joint angle is obtained.

6. The online control method according to claim 5, characterized in that, The kth fuzzy rule of the exoskeleton robot control model for: ; in, Represents the Lagrange multipliers. Here, N represents the kernel function, and N represents the total number of training samples. express The membership degree of the k-th fuzzy rule, This represents the consequent parameters of the exoskeleton robot control model.

7. The online control method according to claim 6, characterized in that, The step of obtaining the expression for the relationship between the controller output torque and the joint angle based on the k-th fuzzy rule of the exoskeleton robot control model includes: The k-th fuzzy rule of the exoskeleton robot control model is obtained by performing a first-order Taylor expansion and taking the derivative. for: ; in, This represents the joint angle of the exoskeleton robot at time T-1 under the k-th fuzzy rule. express The membership degree of the k-th fuzzy rule, Let represent the Lagrange multiplier, and N represent the total number of training samples. The width of the Gaussian function. This indicates that the exoskeleton robot described under the k-th fuzzy rule is... Output torque at any moment This indicates that the exoskeleton robot in the i-th training sample under the k-th fuzzy rule is... Output torque at any moment , , This indicates that the exoskeleton robot under the k-th fuzzy rule is... Output torque at any moment This indicates that the exoskeleton robot under the k-th fuzzy rule is... Output torque at any given moment; in, ,in, This represents the joint angle of the exoskeleton robot at time T-2 under the k-th fuzzy rule. Let represent the joint angle of the exoskeleton robot in the i-th training sample under the k-th fuzzy rule at time T-2. This represents the joint angle of the exoskeleton robot at time T-3 under the k-th fuzzy rule. Let represent the joint angle of the exoskeleton robot in the i-th training sample under the k-th fuzzy rule at time T-3. This represents the joint angle of the exoskeleton robot at time T-4 under the k-th fuzzy rule. Let represent the joint angle of the exoskeleton robot in the i-th training sample under the k-th fuzzy rule at time T-4; Calculating each fuzzy rule after differentiation, the controller output torque under each fuzzy rule is obtained as follows: ; By weighting the controller output torque under each fuzzy rule, the relationship between the controller output torque and the joint angle is obtained as follows: ; in, This represents a preset constant value. Let represent the expected joint angle of the exoskeleton robot at time T under the k-th fuzzy rule. This represents the controller output torque of the exoskeleton robot at time T. This represents the output torque under the k-th fuzzy rule. The coefficients used for weighted summation.

8. An online control device for an exoskeleton robot, characterized in that, include: The modeling module models the exoskeleton robot based on fuzzy rules, resulting in a TS fuzzy model. The rule fusion module fuses the fuzzy rules in the TS fuzzy model and projects the fused TS fuzzy model into a high-dimensional space to obtain the exoskeleton robot control model. The consequent parameter solving module solves for the consequent parameters of the exoskeleton robot control model based on the training data of the exoskeleton robot. The relational expression determination module determines the relational expression between the controller output torque and the joint angle based on the obtained consequent parameters. The exoskeleton robot control module uses the aforementioned relational expression to calculate the controller output torque for the desired joint angle at time T, and controls the controller of the exoskeleton robot at time T based on the calculated controller output torque.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the online control method for the exoskeleton robot as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the online control method for the exoskeleton robot as described in any one of claims 1 to 7.

Citation Information

Patent Citations

  • Whole body exoskeleton carrying and assembling system based on multi-source sensor and control method

    CN113043248A

  • Layered control method for lower limb exoskeleton robot

    CN113855477A