A control method and device of a lower limb rehabilitation robot, a medium and an equipment

Abstract

This invention provides a control method, device, medium, and equipment for a lower limb rehabilitation robot. It relates to the field of robot control technology. The method includes: constructing a two-link model incorporating hip and knee joint movements, and introducing a friction and disturbance model into the two-link model to obtain a dynamic model of the lower limb rehabilitation robot; collecting actual human-machine interaction forces, adjusting admittance and damping parameters in real time based on the time deviation of rehabilitation task execution, thereby converting the position error into a position correction amount and superimposing it onto a preset trajectory to generate a desired trajectory; estimating system disturbances based on the position error to obtain a disturbance estimate; generating a compensation control law based on the position error and the disturbance estimate, and applying the compensation control law to the lower limb rehabilitation robot dynamic model to control the lower limb rehabilitation robot to move along the desired trajectory.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, and in particular to a control method, device, medium and equipment for a lower limb rehabilitation robot. Background Technology

[0002] Lower limb exoskeleton rehabilitation robots, as important devices that can assist patients in gait training and functional reconstruction, have received widespread attention in the field of rehabilitation medicine in recent years. However, most existing lower limb exoskeleton systems adopt a passive control strategy that uses rigid actuators and preset trajectories. This makes it difficult to fully adapt to the patient's active movement intentions during human-machine interaction, easily leading to mechanical traction and even secondary injuries, posing significant safety hazards. Therefore, it is necessary to develop intelligent control methods with good compliance and adaptive capabilities to achieve safe and natural human-machine collaborative movement.

[0003] Current lower limb rehabilitation robots exhibit complex, unmodeled dynamic characteristics during actual operation (such as joint friction and transmission nonlinearity). Furthermore, individual factors such as changes in patient muscle strength and differences in movement patterns, along with external environmental disturbances, further exacerbate the system's uncertainties. These factors make it difficult for traditional control strategies based on precise models to guarantee good tracking performance and interaction safety, severely limiting the reliability of rehabilitation outcomes and clinical applications. Summary of the Invention

[0004] Therefore, it is necessary to provide a control method, device, medium, and equipment for a lower limb rehabilitation robot to address the aforementioned technical problems, thereby solving the issues of robot stiffness, susceptibility to interference, and poor adaptability in the existing technology, and enabling the robot to better cooperate with patients in lower limb rehabilitation training.

[0005] The following technical solution is adopted in this specification: This manual provides a control method for a lower limb rehabilitation robot, including: A two-bar model representing the movement of the thigh around the hip joint and the lower leg around the knee joint is constructed. A friction model representing the motion friction characteristics and an interference model representing the disturbance are introduced into the two-bar model to obtain the dynamic model of the lower limb rehabilitation robot. The actual interaction force between the target rehabilitation subject and the lower limb rehabilitation robot is collected, and the admittance damping parameters are adjusted according to the deviation between the actual execution time of the rehabilitation task and the preset time. Based on the adjusted admittance damping parameters, the difference between the actual interaction force and the preset training rehabilitation force is converted into a position correction amount. The position correction amount is superimposed on the preset trajectory to obtain the desired trajectory of the lower limb rehabilitation robot. The positional error between the desired trajectory and the actual position of the lower limb rehabilitation robot is obtained, and the disturbance is estimated based on the positional error to obtain the disturbance estimate. A compensation control law is generated based on the position error and disturbance estimate. This compensation control law is then applied to the dynamic model of the lower limb rehabilitation robot to control the robot to move along the desired trajectory.

[0006] Furthermore, obtaining the dynamic model of the lower limb rehabilitation robot specifically includes: Obtain the total kinetic energy and total potential energy of the two-bar linkage model: The Lagrangian function of the two-bar model is obtained by subtracting the total potential energy from the total kinetic energy; The derivative of the Lagrangian function of the two-bar model is calculated to obtain the dynamic equations of the thigh motion around the hip joint and the lower leg motion around the knee joint. By incorporating Coulomb friction, viscous friction, and interference terms into the dynamic equations of the thigh's motion around the hip joint and the lower leg's motion around the knee joint, a dynamic model of the lower limb rehabilitation robot is obtained.

[0007] Furthermore, adjusting the admittance damping parameters based on the deviation between the actual execution time and the preset time of the rehabilitation task specifically includes: When the actual execution time is longer than the preset time, it indicates that the actual interaction force has decreased, resulting in motion lag. The motion resistance is reduced by linearly decreasing the admittance damping parameter. When the actual execution time is less than the preset time, it indicates that the actual interaction force increases, causing the motion to be ahead of the target. The admittance damping parameter is linearly increased to increase the motion resistance.

[0008] Furthermore, the method of converting the difference between the actual interaction force and the preset training and rehabilitation force based on the adjusted admittance damping parameters into a position correction amount specifically includes: An admittance model is constructed, which establishes the dynamic relationship between force deviation and position correction through the inertia coefficient, the adjusted admittance damping coefficient, and the environmental stiffness coefficient. The difference between the actual interaction force and the preset training rehabilitation force is used as the input to the admittance model to obtain the position correction amount.

[0009] Furthermore, the estimation of disturbance based on position error to obtain a disturbance estimate specifically includes: Construct a cascaded observer consisting of a primary extended state observer, a PD controller, and a secondary extended state observer; The position error is input into the cascaded observer, and the initial observation value is obtained by making preliminary observations of external disturbances and internal parameter disturbances through the primary extended state observer. The position error is corrected by the PD controller, and the corrected position signal and residual disturbance are input into the secondary extended state observer for secondary observation to obtain the second observation value. By fusing the first and second observations, a disturbance estimate is obtained.

[0010] Furthermore, the generation of the compensation control law based on the position error and disturbance estimate specifically includes: Based on the nonlinear state error feedback law, the position error and the feedback control quantity of the position error differential signal are calculated. The disturbance estimate is used as a feedforward compensation term and superimposed on the feedback control quantity to generate a compensation control law.

[0011] This manual provides a control device for a lower limb rehabilitation robot, including: The model building module is used to construct a two-bar model representing the movement of the thigh around the hip joint and the lower leg around the knee joint. A friction model representing the motion friction characteristics and an interference model representing external uncertain disturbances are introduced into the two-bar model to obtain the dynamic model of the lower limb rehabilitation robot. The desired trajectory generation module is used to collect the actual interaction force between the target rehabilitation object and the lower limb rehabilitation robot, and adjust the admittance damping parameters according to the deviation between the actual execution time of the rehabilitation task and the preset time. Based on the adjusted admittance damping parameters, the difference between the actual interaction force and the preset training rehabilitation force is converted into a position correction amount, and the position correction amount is superimposed on the preset trajectory to obtain the desired trajectory of the lower limb rehabilitation robot. The disturbance estimation module is used to obtain the position error between the desired trajectory and the actual position of the lower limb rehabilitation robot, and to estimate the external uncertain disturbance based on the position error to obtain the disturbance estimate value. The control module is used to generate a compensation control law based on the position error and disturbance estimate, and apply the compensation control law to the dynamic model of the lower limb rehabilitation robot to control the lower limb rehabilitation robot to move along the desired trajectory.

[0012] This specification provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the control method for the lower limb rehabilitation robot described above.

[0013] This specification provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the control method of the lower limb rehabilitation robot described above.

[0014] The above-mentioned technical solutions adopted in this specification can achieve the following beneficial effects: This invention dynamically adjusts the admittance and damping parameters based on the deviation between the actual execution time and the preset time of the rehabilitation task. The force deviation is then converted into a position correction and superimposed on the preset trajectory to generate the desired trajectory. This allows for adaptive adjustment of the admittance control parameters based on individual factors such as changes in patient muscle strength and differences in movement patterns. This ensures that the robot's trajectory always matches the patient's actual movement intention, improving the smoothness and adaptability of human-computer interaction and avoiding incoordination issues caused by fixed parameters and inconsistencies between the trajectory and the patient's intention. Simultaneously, it identifies and quantifies uncertainties caused by individual patient factors and external environmental disturbances. By correcting position deviations and actively canceling various disturbances through a compensatory control law, the robot can consistently track the desired trajectory even in the presence of unmodeled dynamics, individual differences, and external disturbances. This improves the system's anti-interference capability and trajectory tracking accuracy, solving the problems of poor tracking performance and low interaction safety of traditional control strategies based on precise models under the influence of uncertainties. Attached Figure Description

[0015] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0016] Figure 1 This is a schematic diagram of a control method for a lower limb rehabilitation robot provided in this specification; Figure 2 This is a schematic diagram of a two-bar linkage model provided in this specification; Figure 3 This is a schematic diagram of an ADRC structure provided in this specification; Figure 4 This specification provides a schematic diagram of an adaptive admittance cascaded control structure. Figure 5 This is a schematic diagram of an adaptive variation curve of damping parameters provided in this specification. Figure 5 (a) in the figure is a schematic diagram of the changes in damping parameters of the hip and knee joints. Figure 5 (b) in the figure is a schematic diagram of the damping error variation curves of the hip and knee joints. Figure 5 (c) in the figure is a schematic diagram of the variation curve of the coupling damping parameter; Figure 6 This is a schematic diagram of a hip and knee joint angle error tracking curve provided in this specification. Figure 6 (a) in the figure is a schematic diagram of the hip joint angle tracking curve. Figure 6 (b) in the figure is a schematic diagram of the knee joint angle tracking curve. Figure 6 (c) in the figure is a schematic diagram of the hip joint error tracking curve. Figure 6 (d) in the figure is a schematic diagram of the knee joint error tracking curve; Figure 7 This manual provides a schematic diagram of a hip and knee joint velocity tracking curve. Figure 7 (a) in the figure is a schematic diagram of hip joint velocity tracking curves. Figure 7 (b) in the figure is a schematic diagram of the knee joint velocity tracking curve; Figure 8 This is a schematic diagram of the control device for a lower limb rehabilitation robot provided in this specification; Figure 9 This is a schematic diagram of a computer device provided for this specification. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this specification clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments in this specification without creative effort are within the scope of protection of this application.

[0018] To address the challenges of unknown internal model parameters and external disturbances in the adaptive control of lower limb rehabilitation robots, an adaptive admittance control method based on a cascaded observer is proposed. First, the Lagrangian method is used to dynamically model the lower limb rehabilitation robot, quantifying the structural parameters and dynamic characteristics of the exoskeleton. Second, the damping parameters of the system are adaptively adjusted by setting a variable admittance factor, thereby improving the human-machine interaction performance of the lower limb rehabilitation robot under different environments. Finally, the errors caused by changes in internal parameters and external disturbances are dynamically estimated and compensated using a cascaded extended state observer, effectively improving the system's robustness and enabling rapid dynamic adjustment. Simulation results show that compared with traditional adaptive admittance control methods, the tracking accuracy of the hip and knee joints is improved by 5.6%, with the tracking error remaining stable within ±4 degrees. Experimental results demonstrate that this control method effectively improves the system's control accuracy and exhibits good transient response, ensuring the training effect for rehabilitation patients.

[0019] The technical solution of the present invention will now be described in detail with reference to the accompanying drawings.

[0020] Figure 1 This is a schematic diagram of a control method for a lower limb rehabilitation robot provided in this specification, such as... Figure 1 As shown, the method includes the following: S1. Construct a two-bar model to represent the movement of the thigh around the hip joint and the lower leg around the knee joint. Introduce a friction model to represent the motion friction characteristics and an interference model to represent external uncertain disturbances into the two-bar model to obtain the dynamic model of the lower limb rehabilitation robot.

[0021] For example, Figure 2 This is a schematic diagram of a two-bar linkage model provided in this specification, such as... Figure 2 As shown, the lower limb rehabilitation robot's exoskeleton mainly consists of the hip and knee joints. Its movement principle primarily involves the thigh rotating around the hip joint and the lower leg rotating around the knee joint. It is assumed that the mass and length of the thigh strut of the lower limb exoskeleton are respectively... , The mass and length of the lower leg bar are respectively , The distance from the center of mass 'a' of the thigh bar to the center of rotation of the hip joint is... The distance from the center of mass b of the lower leg bar to the center of rotation of the knee joint is . , Let these be the joint variables of the hip and knee joints of the lower limb exoskeleton. Then, based on the Lagrange equation, the total kinetic energy of the two-bar model, and the total potential energy of the two-bar model, the dynamic equations for the thigh and lower leg can be obtained.

[0022] The total kinetic energy of the two-bar linkage model is: ; in, The angular velocity of the hip joint is the rate of rotation of the thigh around the hip joint. The angular velocity of the knee joint is the rate of rotation of the lower leg around the knee joint.

[0023] The total potential energy of the two-bar linkage model is: ; in, This is the acceleration due to gravity.

[0024] The Lagrange equation is: ; in, It is an abbreviation for Lagrange. The total kinetic energy of the system. Let T be the total potential energy of the system, and T be the net external force acting on the joint.

[0025] Substituting the total kinetic energy of the link model and the total potential energy of the two-link model into the Lagrange equation, we obtain the Lagrangian function of the two-link model. for: .

[0026] Based on the Lagrangian function of the two-link model, the dynamic equation of the thigh can be derived. The dynamic equation of the lower leg : .

[0027] .

[0028] in, This is the angular acceleration of the hip joint. This represents the angular acceleration of the knee joint.

[0029] Furthermore, the effects of friction and external interference need to be considered in the actual control system, therefore a friction model needs to be introduced into the lower limb exoskeleton robot. and interference model Friction model for:

[0030] ; in, , It is the Coulomb friction coefficient of the large and small leg robotic arms in the system. , It is the viscous friction coefficient of the large and small leg robotic arms.

[0031] The dynamic equation of the thigh The dynamic equation of the lower leg Based on parameters , and After matrix encapsulation / merging, the resulting dynamic model of the lower limb rehabilitation robot can be represented as: ; in The inertia matrix, For Coriolis force and centrifugal force terms, This is the term related to gravity. This is the frictional torque; This is an uncertain interference term, expressed in N·m; Human-computer interaction torque; The driving torque provided to the driver.

[0032] By constructing a two-bar model representing the movement of the thigh around the hip joint and the lower leg around the knee joint, and by introducing a friction model representing the characteristics of motion friction and a disturbance model representing disturbance, the shortcomings of traditional accurate models that do not consider joint friction, transmission nonlinearity and other unmodeled dynamics are made up for. This allows the dynamic model to better fit the physical characteristics of the robot's actual operation, reduces the impact of the inherent uncertainty of the system on the control effect from the model foundation level, and provides a precise and realistic physical basis for the implementation of subsequent control strategies.

[0033] S2. Collect the actual interaction force between the target rehabilitation object and the lower limb rehabilitation robot, and adjust the admittance damping parameters according to the deviation between the actual execution time of the rehabilitation task and the preset time. Based on the adjusted admittance damping parameters, convert the difference between the actual interaction force and the preset training rehabilitation force into a position correction amount, and superimpose the position correction amount onto the preset trajectory to obtain the desired trajectory of the lower limb rehabilitation robot.

[0034] For example, rehabilitation patients need to ensure compliance with the lower limb exoskeleton during rehabilitation training to improve safety mechanisms. In designing flexible contact, admittance control is generally used as a compliance control strategy. Admittance control is a force-based compliance control method that achieves compliance by establishing a relationship with position. The general expression for admittance control is:

[0035] ; in, , , These are the inertia coefficient, damping coefficient, and environmental stiffness coefficient; It is the current position of the object. For the desired position; , These are the actual interaction force between the human lower limb and the exoskeleton of the lower limb rehabilitation robot, and the preset training rehabilitation force, respectively. For actual speed, For the desired speed, For actual acceleration, The desired acceleration.

[0036] For example, Fitts's Law, an important law in psychology and human-computer interaction, describes the relationship between the time required for a human to move quickly and precisely to a target and the distance and size of the target. The time required to move to the target... From the target distance Determined by both the size of the target and its objective, its formula is expressed as:

[0037] ; in The time to complete the action, The distance from the starting point to the center of the target. The effective width of the target in the direction of motion. , is the experimental fitting constant.

[0038] For example, currently, the admittance parameters for lower limb exoskeleton admittance control are fixed. However, the muscle strength exhibited by rehabilitation patients varies depending on the stage of rehabilitation and the degree of muscle strain, and the damping coefficient also changes accordingly. Therefore, to ensure the normal operation of the lower limb exoskeleton, improve the fit between the rehabilitation robot and the rehabilitation patient, and thus ensure the compliance of the rehabilitation exoskeleton, this embodiment adopts the Fitz adaptive admittance control principle. According to the Fitz law principle, the admittance damping parameter formula is as follows:

[0039] ; in The modified damping parameters, This is the initial damping value. It is a linear variable admittance factor.

[0040] To ensure that the empirical parameters of Fitz's law formula remain constant when the admittance changes, Conditions to be met: ; in, The preset duration for rehabilitation tasks.

[0041] In a certain rehabilitation task, the patient moves the walking aid from an initial position to an end point with stable force. The duration of this task is... When the actual task time hour, ;when This indicates that the reduced interaction force caused the walker to arrive after the set time, at which point the damping value needs to be reduced; when This indicates that the increased interaction force causes the walker to arrive before the set time, at which point the damping value needs to be increased. Based on this, an adaptive damping variation formula is established:

[0042] ; in, It is a linear rate of change, and .

[0043] Based on Fitz's law, the admittance-damping parameter formula, and the adaptive damping variation formula, the linear variable admittance factor can be obtained. With linear rate of change The relationship is as follows: .

[0044] The specific implementation process of converting the difference between the actual interaction force and the preset training rehabilitation force into a position correction quantity based on the adjusted admittance damping parameters and superimposing it onto the preset trajectory to obtain the desired trajectory is as follows: First, the admittance damping parameters obtained through adaptive adjustment using Fitz's law are substituted into the general expression of admittance control. Simultaneously, the preset inertia coefficient and environmental stiffness coefficient are fixed. The actual interaction force between the target rehabilitation object and the lower limb rehabilitation robot, the preset training rehabilitation force, and the robot's real-time current position, current velocity, and current acceleration are collected. All the above known parameters and physical quantities are substituted into the admittance control expression. Using the desired position, desired velocity, and desired acceleration as solution variables, the expression is obtained by solving the dynamic equations. The expected position correction value is obtained by mapping the difference between the actual interaction force and the preset training rehabilitation force through the admittance model. Then, the standardized rehabilitation training trajectory preset by the lower limb rehabilitation robot for the target rehabilitation object is retrieved. The preset trajectory includes the standard position coordinate sequence of the hip and knee joints at each stage of rehabilitation training. The obtained position correction value is superimposed point by point onto the standard position coordinates corresponding to the preset trajectory according to the time node and joint dimension of rehabilitation training. The position of each node of the preset trajectory is corrected. The trajectory containing the new position coordinate sequence of the hip and knee joints after correction is the expected trajectory of the lower limb rehabilitation robot adapted to the current movement state and muscle strength level of the target rehabilitation object.

[0045] S3. Obtain the positional error between the desired trajectory and the actual position of the lower limb rehabilitation robot, estimate the disturbance based on the positional error, and obtain the disturbance estimate.

[0046] S3.1 Construct a cascaded observer (OBR-ESO) consisting of a primary extended state observer (ESO1), a PD controller, and a secondary extended state observer (ESO2).

[0047] For example, ADRC (Active Disturbance Rejection Control) is a control theory whose core principle is to achieve precise control of the controlled object by actively estimating and compensating for various disturbances in the system, without relying on an accurate mathematical model. Figure 3 This specification provides a schematic diagram of an ADRC structure, such as... Figure 3As shown, its core components consist of four parts: an Extended State Observer (ESO), a Tracking Differentiator (TD), a Nonlinear State Error Feedback (NLSEF), and a disturbance compensation device. In ADRC, the TD generates a smooth tracking signal and its differential signal, avoiding system oscillations caused by sudden input changes. The ESO estimates and processes the system's state variables and their internal and external disturbances in real time by observing the system's input and output. The NLSEF quickly eliminates errors and compensates for internal and external disturbances by comparing the tracking signal output by the TD with the actual state signal estimated by the ESO. This method treats model inaccuracies and coupling elements as internal disturbances of the system, combining them with external disturbances to form the total system disturbance. This disturbance is then expanded into an additional state variable, which is estimated by the ESO along with other state variables. Therefore, ADRC achieves the transformation of the controlled object into a linearized and deterministic form.

[0048] In the movement process of lower limb rehabilitation exoskeleton robots, there are high-order unknown disturbances with multivariable coupling, and the low-order state observer has low accuracy in observing these disturbances. To address this, this embodiment designs a cascaded observer (OBR-ESO) based on two ESOs to enhance the disturbance observation capability. Figure 4 This specification provides a schematic diagram of an adaptive admittance cascaded control structure, as shown below. Figure 4 As shown, the specific design steps are as follows: First, a primary observer ESO1 is designed to observe generalized external disturbances and internal system disturbances. Second, a proportional-derivative (PD) controller is designed to perform early corrections before position errors occur, and the corrected position signal and residual disturbances are estimated using a secondary observer ESO2. Finally, the observation values ​​of the primary observer ESO1 and the secondary observer ESO2 are fed back into the system as compensation terms, thereby achieving the purpose of dual observation.

[0049] A single-input, single-output nonlinear system model is performed on the lower limb rehabilitation exoskeleton system with perturbations. The fundamental relationships between the system output, input, and internal and external perturbations are defined to provide the original system model for subsequent observer design. The lower limb rehabilitation exoskeleton system is assumed to be a single-input, single-output nonlinear system with perturbations, and its input-output equations are as follows: ; in, The system output corresponds to the actual position signals of the hip / knee joints of the lower limb rehabilitation robot. It is represented as a nonlinear function of generalized external disturbances and internal system disturbances; For unknown external disturbances; This is the system control input, corresponding to the control torque / voltage signal that drives the robot's motion; The input coefficients represent the nth-order dynamic changes in the robot's joint positions, determined by the combined effect of all disturbances and the driving force of the control input. This formula simplifies the complex multi-disturbance nonlinear system into a disturbance term plus a control term, forming the basis for subsequent expansion of the disturbances into state variables and the design of the observer. This input-output equation provides the original system model for observer design; and by making the disturbance function differentiable and introducing extended state variables, the state-space description of the system is completed, yielding the observer's dynamic equations, thus laying the model foundation for the specific design of ESO1 and ESO2.

[0050] Assumption Differentiable, it is defined as the sum of internal and external disturbances. ,Right now Set the system output position to At the same time, an extended state is introduced. By treating the unknown disturbances of the system as an observable state variable, the original high-order nonlinear system is transformed into a system containing... The linear state-space form of the state variables. Based on the above state-space description, the general dynamic equation of the observer can be obtained:

[0051] ; in, To simulate the state vector of the system, ; , , These are the system's state matrix, input matrix, and output matrix, respectively. The observer outputs a feedback matrix, which determines the speed and stability of error correction. This represents the output vector of the simulation system. This equation lays the model foundation for the specific design of ESO1 and ESO2.

[0052] S3.2 Input the position error into the cascaded observer, and use the primary extended state observer to make preliminary observations of external disturbances and internal parameter disturbances to obtain the first observation value.

[0053] make Then the primary observer ESO1 is: ; ; in, , , These are the inertia coefficient, damping coefficient, and stiffness coefficient, respectively. This reflects the system's resistance to acceleration. This reflects the system's ability to impede speed. This reflects the system's ability to correct positional deviations.

[0054] Setting the inertia coefficient for the robot system =1, Damping coefficient =20, Stiffness coefficient =10, and define the gain vector. By substituting the above parameters into the matrix ,get:

[0055] .

[0056] The matrix of the parameters to be substituted The characteristic root equations are: Solving this characteristic root equation yields the gain parameter of ESO1. The final equation of state for ESO1 is determined as follows: By inputting the position error into ESO1 based on this equation, preliminary observations of external disturbances and internal parameter disturbances can be completed, and the first observation value can be output.

[0057] S3.3 The position error is corrected by the PD controller, and the corrected position signal and residual disturbance are input into the secondary extended state observer for secondary observation to obtain the second observation value.

[0058] After the primary extended state observer ESO1 completes the initial observation of external disturbances and internal parameter disturbances and outputs the first observation value, there will still be some residual disturbances that have not been fully observed. Furthermore, the position error still needs to be further accurately corrected after the initial observation compensation. Therefore, the original position error is corrected in real time by the PD controller, and then the corrected position signal and the residual disturbance are input together into the secondary extended state observer ESO2 for a second accurate observation, thereby improving the overall accuracy of disturbance observation.

[0059] Similarly, continuing the state variable definition logic of ESO1, let Then the secondary observer ESO2 is: ; ; in, This is the gain parameter used for proportional control in the PD controller. It is mainly used to quickly eliminate the current error, thereby affecting the response speed of the control system to the current deviation. The gain parameter of the PD controller for derivative control is mainly used to suppress rapid changes in deviation to enhance stability, thus determining the response speed of the control system to the rate of change of deviation.

[0060] To ensure consistency in observer design and system adaptability, the same system parameters as ESO1 were used. =1,B =20, K =100, Construct the system matrix of ESO2 and derive the closed-loop characteristic equation: .

[0061] After simplification and derivation, the closed-loop characteristic equation of the zero-free second-order system ESO2 can be obtained as follows: .

[0062] To achieve accurate observation of residual disturbances and stable system response, appropriate PD controller parameters are selected. =5, =0, substituting into the above closed-loop characteristic equation, we obtain the key parameters of ESO2: , .

[0063] Substituting all the solved parameters into the system matrix and state equations of ESO2, the final dynamic equations of ESO2 are determined, which are as follows: .

[0064] After constructing and tuning the dynamic equations of ESO2, the position signal corrected by the PD controller and the residual disturbances in the system not observed by ESO1 are used as inputs and substituted into the dynamic equations of ESO2 for iterative calculation. This allows for a second, accurate observation of the system disturbance and outputs the second observation value. By fusing the first and second observation values, the disturbance estimate is obtained.

[0065] S4. Generate a compensation control law based on the position error and disturbance estimate, and apply the compensation control law to the dynamic model of the lower limb rehabilitation robot to control the lower limb rehabilitation robot to move along the desired trajectory.

[0066] For example, a compensation control law is generated based on the position error and disturbance estimate, and then applied to the dynamic model of the lower limb rehabilitation robot to control the lower limb rehabilitation robot to move along the desired trajectory. This step is based on the position error and disturbance estimate obtained in S3, and combines the nonlinear state error feedback (NLSEF) theory in ADRC to generate the compensation control law.

[0067] The core function of NLSEF in ADRC is to quickly eliminate errors by comparing the tracking signal with the actual state signal estimated by the observer. In this step, this theory is used to perform nonlinear calculations on the position error between the desired trajectory and the actual position obtained in S3, as well as the differential signal of the position error, to obtain a feedback control quantity that can eliminate the position error, providing a foundation for the subsequent generation of the control law. The disturbance estimate obtained in S3 is superimposed on the feedback control quantity as a feedforward compensation term to generate a compensation control law. Applying this compensation control law to the lower limb rehabilitation robot dynamic model constructed in S1, the robot can be precisely controlled by the dual effects of feedforward compensation of disturbance and feedback correction of position error, making it strictly follow the desired trajectory generated in S2.

[0068] Furthermore, this manual also provides simulation results and analysis, specifically including: Lower limb rehabilitation exoskeleton robots are characterized by unknown model parameters, strong coupling, and susceptibility to interference. To address these issues, this invention proposes a cascaded adaptive admittance control strategy based on active disturbance rejection. To verify the effectiveness of the proposed method, a system simulation model was built on the Matlab / Simulink platform, using the hip and knee joints as the research objects of the lower limb exoskeleton rehabilitation robot. The actual torque of the rehabilitation patient was measured by force sensors and used as the input to the control system along with the set desired torque. Angle and angular velocity were used as the outputs of the control system. Finally, a simulation model was built in the Simulink system, simulation parameters were set, and the control system was jointly simulated. The simulation results were then analyzed systematically.

[0069] Analysis of the adaptive process of damping parameters.

[0070] This invention employs an adaptive admittance control strategy based on Fitz's law. By setting an adaptive admittance factor, the damping coefficient of the system is altered, thereby improving the fit between the lower limb rehabilitation exoskeleton robot and the rehabilitation patient, achieving compliance of the rehabilitation exoskeleton robot during human-machine interaction. Let the desired damping coefficient of the hip joint be... The expected damping coefficient of the knee joint According to the simulation results, the damping coefficient of the hip joint approaches the expected value at 1.426 seconds, and the damping coefficient of the knee joint approaches the expected value at 2.345 seconds. Furthermore, the error eventually steadily approaches 0, demonstrating good convergence. The simulation results are as follows: Figure 5 As shown.

[0071] Follow-up control experiment.

[0072] During actual rehabilitation training, there is inevitably a certain error between the actual and expected motion angles of the hip and knee joints of patients. To ensure the safety of patients during rehabilitation training, the angle tracking error of the hip and knee joints needs to be controlled within a certain range to avoid accidents. Therefore, this invention, based on adaptive admittance control, introduces a cascaded observer to dual-observe both internal positional errors and external disturbances, thereby improving the tracking accuracy of the hip and knee joints. Finally, appropriate admittance parameters are selected based on the patient's body shape and degree of motor dysfunction to simulate and analyze the hip and knee joint angle tracking and error tracking. The simulation results are as follows: Figure 6 As shown. Figure 6 Experimental results show that the hip and knee joint angle tracking performance is good. The root mean square error (RMSE) for hip joint angle tracking is 2.94°, with a maximum error of 6.62°; the root mean square error (RMSE) for knee joint angle tracking is 2.97°, with a maximum error of 6.59°. Furthermore, the tracking errors remain within ±4°, and the maximum error remains within 7°. Compared to traditional adaptive admittance control, its tracking accuracy is improved by 5.6%. Simulation results show that the error fluctuation range of this method fully meets the requirements of lower limb rehabilitation exoskeleton robots in the rehabilitation training process.

[0073] Admittance velocity tracking experiment based on ADRC.

[0074] During actual training, rehabilitation patients' lower limb exoskeletons are easily affected by uncertain factors such as external obstacles, uneven ground surfaces, internal friction, and elastic deformation of mechanical structures. By employing an ADRC-based adaptive admittance control method, after a brief adjustment period in the initial stage of movement when subjected to external disturbances, the actual velocity can effectively track the admittance velocity, with a tracking accuracy within ±1.6 deg / s. This effectively improves the system's anti-interference capability and ensures the compliance of the movement process. The hip and knee joint admittance velocity tracking curves of this invention, using ADRC-based adaptive admittance control, are shown below. Figure 7 As shown.

[0075] Analysis of experimental results.

[0076] Lower limb rehabilitative exoskeletons have a positive effect on lower limb function rehabilitation in stroke patients and other populations. Meta-analysis shows that lower limb rehabilitative exoskeletons have a positive effect on lower limb function rehabilitation in stroke patients and other populations. Related meta-analyses show that exoskeleton robots can significantly improve the lower limb motor function FMA-LE score, balance function BBS score, and stand-up-walk test TUG score in stroke patients. A controlled experiment was conducted, selecting thirty healthy subjects of different ages and randomly dividing them into three groups: a traditional admittance group, an adaptive admittance group, and an ADRC-adaptive admittance group for training. After a period of training, scores were given based on the Fugl-Meyer Assessment (FMA), the 10-meter walk test (10MWT), and self-satisfaction. The scoring results are shown in Table 1. According to the experimental results, compared with the previous two methods, the method proposed in this invention has significant advantages in all three evaluation indicators.

[0077] Table 1. Trial Satisfaction Rating Table In summary, to address the problems of unknown model parameters, inaccurate parameter identification, and unknown disturbances in trajectory tracking control of lower limb exoskeleton rehabilitation robots, this invention proposes an adaptive admittance control strategy based on a cascaded observer. First, considering the different stiffness coefficients exhibited by patients at different rehabilitation stages due to varying muscle strength, an adaptive admittance control principle based on Fitz's law is adopted. By introducing a linear variable admittance factor to adaptively adjust the stiffness coefficient of the lower limb rehabilitation robot, its adaptive capability is improved, thereby ensuring the robot's compliance. Second, a PD controller is introduced at the control input to correct for position errors before they occur, thus improving system performance. Finally, a cascaded state observer is designed to dual-observe system disturbances and position errors, thereby improving the system's anti-interference capability and control accuracy. Simulation experiments on the Matlab platform demonstrate that the proposed strategy achieves good tracking accuracy and anti-interference capability while ensuring compliance.

[0078] The control device for the lower limb rehabilitation robot provided by the present invention is described below. The control device for the lower limb rehabilitation robot described below can be referred to in correspondence with the control method for the lower limb rehabilitation robot described above.

[0079] Figure 8 This is a schematic diagram of the control device for a lower limb rehabilitation robot provided by the present invention. For example, please refer to... Figure 8 As shown, the control device for this lower limb rehabilitation robot may include: The modeling module is used to perform dynamic modeling of the lower limb exoskeleton rehabilitation robot using the Lagrange method, and to establish a dynamic model that includes system parameters, friction and disturbances; The adaptive module is used to design an adaptive admittance control law based on Fitz's law. By setting a variable admittance factor, it adaptively adjusts the damping parameters in the admittance model to adapt to the human-computer interaction requirements in different environments. The dynamic estimation module is used to design a cascaded extended state observer to dynamically estimate the total error caused by changes in internal system parameters and external disturbances, and output a compensation signal. The drive module is used to superimpose the control command output by the adaptive admittance control law with the compensation signal output by the cascaded extended state observer to obtain the final control quantity, which drives the lower limb exoskeleton rehabilitation robot to move.

[0080] Specific limitations regarding the control device of the lower limb rehabilitation robot can be found in the above description of the control limitations for the lower limb rehabilitation robot, and will not be repeated here. Each module in the control device of the lower limb rehabilitation robot can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0081] This specification also provides a computer-readable storage medium storing a computer program that can be used to execute the above-described... Figure 1 The control method for the provided lower limb rehabilitation robot.

[0082] This instruction manual also provides Figure 9 The schematic diagram of the computer device shown is as follows: Figure 9 At the hardware level, the computer device includes a processor, internal bus, network interface, memory, and non-volatile memory, and may also include other hardware required for business operations. The processor reads the corresponding computer program from the non-volatile memory into memory and then runs it to achieve the above-mentioned functions. Figure 1 The control method for the provided lower limb rehabilitation robot.

[0083] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0084] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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 specification.

Claims

1. A control method of a lower extremity rehabilitation robot characterized by comprising: include: A two-bar model representing the movement of the thigh around the hip joint and the lower leg around the knee joint was constructed. A friction model representing the motion friction characteristics and an interference model representing the disturbance were introduced into the two-bar model to obtain the dynamic model of the lower limb rehabilitation robot. The actual interaction force between the target rehabilitation subject and the lower limb rehabilitation robot is collected, and the admittance damping parameters are adjusted according to the deviation between the actual execution time of the rehabilitation task and the preset time. Based on the adjusted admittance damping parameters, the difference between the actual interaction force and the preset training rehabilitation force is converted into a position correction amount. The position correction amount is then superimposed on the preset trajectory to obtain the desired trajectory of the lower limb rehabilitation robot. The positional error between the desired trajectory and the actual position of the lower limb rehabilitation robot is obtained, and the disturbance is estimated based on the positional error to obtain the disturbance estimate. A compensation control law is generated based on the position error and disturbance estimate. This compensation control law is then applied to the dynamic model of the lower limb rehabilitation robot to control the robot to move along the desired trajectory.

2. The control method of the lower extremity rehabilitation robot according to claim 1, wherein The process of obtaining the dynamic model of the lower limb rehabilitation robot specifically includes: Obtain the total kinetic energy and total potential energy of the two-bar linkage model: The Lagrangian function of the two-bar model is obtained by subtracting the total potential energy from the total kinetic energy; The derivative of the Lagrangian function of the two-bar model is calculated to obtain the dynamic equations of the thigh motion around the hip joint and the lower leg motion around the knee joint. By incorporating Coulomb friction, viscous friction, and disturbance terms of the lower limb rehabilitation robot into the dynamic equations of the thigh's motion around the hip joint and the lower leg's motion around the knee joint, a dynamic model of the lower limb rehabilitation robot is obtained.

3. The control method of the lower extremity rehabilitation robot according to claim 1, wherein The adjustment of admittance damping parameters based on the deviation between the actual execution time and the preset time of the rehabilitation task specifically includes: When the actual execution time is longer than the preset time, it indicates that the actual interaction force has decreased, resulting in motion lag. The motion resistance is reduced by linearly decreasing the admittance damping parameter. When the actual execution time is less than the preset time, it indicates that the actual interaction force increases, causing the motion to be ahead of the target. The admittance damping parameter is linearly increased to increase the motion resistance.

4. The control method of the lower extremity rehabilitation robot according to claim 1, wherein The adjustment of admittance damping parameters converts the difference between the actual interaction force and the preset training and rehabilitation force into a position correction amount, specifically including: An admittance model is constructed, which establishes the dynamic relationship between force deviation and position correction through the inertia coefficient, the adjusted admittance damping coefficient, and the environmental stiffness coefficient. The difference between the actual interaction force and the preset training rehabilitation force is used as the input to the admittance model to obtain the position correction amount.

5. The control method of the lower extremity rehabilitation robot according to claim 1, wherein The estimation of disturbance based on position error to obtain a disturbance estimate specifically includes: Construct a cascaded observer consisting of a primary extended state observer, a PD controller, and a secondary extended state observer; The position error is input into the cascaded observer, and the initial extended state observer is used to make preliminary observations of external disturbances and internal parameter disturbances to obtain the first observation value. The position error is corrected by the PD controller, and the corrected position signal and residual disturbance are input into the secondary extended state observer for secondary observation to obtain the second observation value. By fusing the first and second observations, a disturbance estimate is obtained.

6. The control method for the lower limb rehabilitation robot as described in claim 1, characterized in that, The generation of the compensation control law based on the position error and disturbance estimate specifically includes: Based on the nonlinear state error feedback law, the position error and the feedback control quantity of the position error differential signal are calculated. The disturbance estimate is used as a feedforward compensation term and superimposed with the feedback control quantity to generate a compensation control law.

7. A control device for a lower limb rehabilitation robot, characterized in that, include: The model building module is used to construct a two-bar model representing the movement of the thigh around the hip joint and the lower leg around the knee joint. A friction model representing the motion friction characteristics and an interference model representing external uncertain disturbances are introduced into the two-bar model to obtain the dynamic model of the lower limb rehabilitation robot. The expected trajectory generation module is used to collect the actual interaction force between the target rehabilitation object and the lower limb rehabilitation robot, and adjust the admittance damping parameters according to the deviation between the actual execution time of the rehabilitation task and the preset time. Based on the adjusted admittance damping parameters, the difference between the actual interaction force and the preset training rehabilitation force is converted into a position correction amount. The position correction amount is then superimposed on the preset trajectory to obtain the desired trajectory of the lower limb rehabilitation robot. The disturbance estimation module is used to obtain the position error between the desired trajectory and the actual position of the lower limb rehabilitation robot, and to estimate the external uncertain disturbance based on the position error to obtain the disturbance estimate value. The control module is used to generate a compensation control law based on the position error and disturbance estimate, and apply the compensation control law to the dynamic model of the lower limb rehabilitation robot to control the lower limb rehabilitation robot to move along the desired trajectory.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the control method for the lower limb rehabilitation robot as described in any one of claims 1 to 6.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the control method for the lower limb rehabilitation robot as described in any one of claims 1 to 6.

Images