Control method of lower limb rehabilitation exoskeleton robot based on super-twisting terminal sliding mode

Abstract

A control method for a lower limb rehabilitation exoskeleton robot includes: establishing a single-leg lower limb exoskeleton dynamic model using motor output torque as the system control input signal, and transforming it into a second-order differential equation of the lower limb exoskeleton robot using the Euler-Lagrange method to complete system modeling; designing a non-singular terminal sliding surface based on the lower limb rehabilitation exoskeleton robot model; designing a super-torsional sliding mode control algorithm based on the constructed non-singular terminal sliding surface for the established lower limb rehabilitation exoskeleton robot dynamic model; furthermore, initializing the lower limb exoskeleton robot is required, and the exoskeleton's motion mode is determined through a signal input device; simultaneously, the signals collected by the joint motor encoders are input to the super-torsional terminal sliding mode controller, which calculates the motor output torque that has a good tracking effect on the desired gait trajectory based on the information returned by the sensors and according to the current exoskeleton joint angle and the set joint angle.

Description

Technical Field

[0001] This invention belongs to the field of robot control technology, and more specifically, relates to a trajectory tracking control method for a lower limb rehabilitation exoskeleton robot based on a super-twisted terminal sliding mode. Background Technology

[0002] Among the elderly population, cerebrovascular and neurological diseases are prevalent, with stroke patients being particularly prominent. Even among survivors, most suffer from hemiplegic gait or paralysis. Furthermore, the average age of stroke patients is increasingly younger, with many young adults (20-54 years old) facing disabilities due to musculoskeletal and neurological disorders. However, there is currently a severe shortage of rehabilitation physicians in my country, leaving many patients unable to receive effective and timely rehabilitation treatment.

[0003] Furthermore, with the rapid development of various disciplines such as traditional mechanics, sensor technology, biomedicine, life sciences, and computer control theory, robotics technology in the medical rehabilitation field has ushered in a period of rapid development. Among these, exoskeleton robots, as a new type of robot, have been extensively developed and applied in military, civilian, and medical rehabilitation fields. Wearable lower limb rehabilitation exoskeleton robots are gradually replacing traditional rehabilitation training equipment and have been successfully applied in rehabilitation training for people with disabilities and the elderly. As an intelligent assistive, orthotic, and rehabilitation device, lower limb rehabilitation exoskeleton robots effectively help improve the walking ability of the elderly and people with disabilities, and are of great significance in alleviating my country's aging population problem.

[0004] Because patients have limited mobility in the early stages of rehabilitation, the passive control effect of lower limb rehabilitation exoskeleton robots plays a crucial role in the early rehabilitation outcomes. However, during use, lower limb rehabilitation exoskeleton robots are susceptible to interference from various factors such as internal parameter perturbations, external environment, and improper user operation, making it difficult to guarantee their safety and reliability, and even posing a potential risk of secondary injury to patients. To address these issues and better utilize the performance of lower limb rehabilitation exoskeleton robots, a novel control method is urgently needed for their passive control mode. This method would enable real-time control of motor inputs and adjustment of joint angles to achieve the desired recovery effect. Common methods include PID control, adaptive control, sliding mode control, fuzzy logic and neural network control, and robust control. Among these methods, sliding mode control is widely used in robot control due to its superior robustness and ease of implementation. However, sliding mode control also has drawbacks: 1. It exhibits chattering during control, resulting in significant motor wear; 2. It lacks strong robustness in the control arrival phase. Therefore, in the application of sliding mode control, how to retain the advantages of sliding mode control while overcoming the above-mentioned disadvantages has become an urgent problem to be solved. Summary of the Invention

[0005] This invention addresses the high-precision passive control requirements of current lower limb rehabilitation exoskeleton robots, overcoming the problems of jitter in control input and insufficient tracking accuracy in existing technologies. It provides a control method for lower limb rehabilitation exoskeleton robots based on super-torsional end-effector sliding mode. This invention offers a kinematic modeling approach for lower limb rehabilitation exoskeleton robots and designs a novel control method based on super-torsional end-effector sliding mode.

[0006] To achieve the above objectives, the technical solution of this invention is: a control method for a lower limb rehabilitation exoskeleton robot based on a super-twisted end sliding mode. The lower limb exoskeleton robot includes relative encoders installed on the knee and ankle joints of the exoskeleton for collecting and calculating the corresponding joint motion angles and angular accelerations. The main steps of the control method include the following:

[0007] 1. Using the motor output torque as the system control input signal, establish a dynamic model of a single-leg lower limb exoskeleton, and transform it into a second-order differential equation for the lower limb exoskeleton robot using the Euler-Lagrange method to complete the system modeling;

[0008] 2. Design a non-singular terminal sliding surface based on a lower limb rehabilitation exoskeleton robot model.

[0009] 3. Based on the established dynamic model of the lower limb rehabilitation exoskeleton robot, a super-twisted sliding mode control algorithm is designed.

[0010] 4. Initialize the lower limb exoskeleton robot by determining the exoskeleton's motion mode through a signal input device.

[0011] 5. Input the signal collected by the joint motor encoder to the super-twisted terminal sliding mode controller. The controller will calculate the motor output torque that has a good tracking effect on the desired gait trajectory based on the information returned by the sensor and according to the current exoskeleton joint angle and the set joint angle.

[0012] Preferably, the exoskeleton's movement mode is set to walking in step 4.

[0013] This invention addresses the gait trajectory tracking problem in the passive control mode of a lower limb rehabilitation exoskeleton robot, and presents a design method for a super-twisted terminal sliding mode control law that ensures trajectory tracking performance, exhibiting high control accuracy, strong robustness, and anti-interference capability.

[0014] The beneficial effects of this invention are mainly reflected in the following aspects: 1. A novel modeling method for lower limb exoskeleton robots with parameter uncertainty and disturbance terms is proposed, realizing passive control of lower limb rehabilitation exoskeleton robots under the conditions of model parameter uncertainty and noise interference, thereby improving model accuracy; 2. A novel super-twisted sliding mode control algorithm is designed, which realizes the finite-time convergence of joint angle tracking error and effectively suppresses the chattering phenomenon commonly found in sliding mode control, thereby improving control quality. Attached Figure Description

[0015] Figure 1 This is a structural diagram of the lower limb rehabilitation exoskeleton robot of the present invention.

[0016] Figure 2 This is a block diagram of the control structure of the lower limb rehabilitation exoskeleton robot of the present invention.

[0017] Figure 3 This is a schematic diagram of the gait trajectory of the lower limb rehabilitation exoskeleton robot of the present invention.

[0018] Figure 4 This is a gait tracking effect diagram of the thigh joint of a lower limb rehabilitation exoskeleton robot.

[0019] Figure 5 This is a gait tracking effect diagram of the lower limb rehabilitation exoskeleton robot's lower leg joint.

[0020] Figure 6 This is a schematic diagram of the thigh joint tracking error of a lower limb rehabilitation exoskeleton robot.

[0021] Figure 7 This is a schematic diagram of the tracking error of the lower leg joint in a lower limb rehabilitation exoskeleton robot.

[0022] Figure 8 This is a flowchart of the method of the present invention. Detailed Implementation

[0023] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0024] The control method for a lower limb rehabilitation exoskeleton robot based on super-twisted terminal sliding mode of the present invention includes the following steps:

[0025] 1. Using the motor output torque as the system control input signal, establish a dynamic model of a single-leg lower limb exoskeleton, and transform it into a second-order differential equation for the lower limb exoskeleton robot using the Euler-Lagrange method to complete the system modeling, including the following steps:

[0026] 1.1 First, according to Figure 1 We can obtain the equations relating the joint angles and positions of the lower limb rehabilitation exoskeleton robot:

[0027]

[0028] The total kinetic energy and total potential energy of the system can be obtained as follows:

[0029]

[0030] In the lower limb exoskeleton mechanical leg model, m1 is the mass of the thigh rod, l1 is the length, and l c1 m1 is the distance from the pivot to the center of mass of the thigh; m2 is the mass of the lower leg bar; l2 is the length; l c2 Let θ1 be the distance from the pivot to the center of mass of the lower leg, and θ2 be the angular positions of the hip and knee joints, respectively. Let p1(x1, y1) and p2(x2, y2) be the coordinates of the centers of mass of the thigh and lower leg, respectively.

[0031] 1.2 Dynamic modeling using the Lagrange method. The Lagrange variables are constructed as follows:

[0032] L = E k -E p (3)

[0033] Substituting this into the following partial differential equation:

[0034]

[0035] The Lagrange equations for the thigh and calf are obtained as follows:

[0036]

[0037]

[0038] 1.3 Based on the above methods, we have derived the standard form of the second-order differential equation for the dynamic model of the lower limb exoskeleton robot as follows:

[0039]

[0040] Where p represents the position vector of the exoskeleton joint, d is the set of external perturbations, and M... p C p and G p These represent the inertia matrix, Coriolis force and centrifugal force matrices, and gravitational torque vector, respectively. Furthermore, M... p C p and G p Specifically, it is expressed as follows:

[0041]

[0042] Where τ is the control input torque that needs to be designed in the second step.

[0043] 2. To achieve high gait trajectory tracking accuracy and strong robustness against parameter uncertainties and unmodeled dynamics, this patent designs a non-singular terminal sliding surface based on a lower limb rehabilitation exoskeleton robot model. The specific steps are as follows:

[0044] 2.1 Define the error variable:

[0045]

[0046] 2.2. Taking the derivative with respect to the error variable, we obtain the expression for the first derivative of the error variable:

[0047]

[0048] 2.3 Based on the above technical solution, the non-singular fast terminal sliding surface is constructed as follows:

[0049]

[0050] Based on the designed non-singular fast terminal sliding surface, we control the system state to move onto the sliding surface s. When the system state moves onto and remains on the sliding surface s, the system tracking error e converges to zero. This specification gives the following definition of the power of a column matrix: e ρ =[e1 ρ e2 ρ ].

[0051] Define Lyapunov functions Taking the first derivative of this function, we get:

[0052]

[0053] 3. For the established dynamic model of the lower limb rehabilitation exoskeleton robot, based on the constructed non-singular terminal sliding surface, specifically including:

[0054] 3.1 When designing a sliding mode control law, the first step is to design an equivalent control law based on the designed sliding surface. This begins by assuming the system parameter uncertainties and external disturbances are d = [d1 d2]. T =[0 0] T Based on the sliding surface s designed in the above steps, the equivalent control law is designed as follows:

[0055]

[0056] in

[0057] 3.2 Following the steps above, based on the sliding mode control principle and the super-twist algorithm, the super-twist reaching law is designed as follows:

[0058]

[0059] Where k1 and k2 are positive constants.

[0060] 3.3 Finally, combining the two control laws, a super-twisted sliding mode control algorithm is designed, and the final control input is as follows:

[0061] τ=τ1+τ2 (14)

[0062] 3.4 Substituting the obtained control input τ into the Lyapunov function, we get:

[0063]

[0064] When the system disturbance d is 0, we can obtain:

[0065]

[0066] Based on k1 and k2 > 0, we can conclude that the first derivative of the Lyapunov function is less than zero, and the system state converges in finite time.

[0067] When the system disturbance d is not 0, then

[0068]

[0069] It can be obtained that in k1-ds -1 / 2 When sgn(s) > 0, the system is stable, and the sliding variable s can converge to the region (dk1) in a finite time. -1 ) 2 Within, that is, the system tracking error e can converge to the region (dk1). -1 ) 2 / γ Inside.

[0070] Therefore, given a reasonable approach control gain, it can be ensured that the first derivative of the Lyapunov function is less than zero and the system state is stable; moreover, the designed super-twisted approach law is a continuous approach law, which can effectively alleviate chattering in the control input.

[0071] 4. Initialization of the lower limb exoskeleton robot: The exoskeleton's motion pattern is determined through signal input devices. To better accomplish the task of lower limb rehabilitation, the lower limb exoskeleton's motion pattern is usually given as a walking model.

[0072] 5. The signals collected by the joint motor encoders are input to the ultra-twisted end-effector sliding mode controller. Based on the information returned by the sensors and according to the current and set joint angles of the exoskeleton, the controller calculates the motor output torque that provides good tracking of the desired gait trajectory. The device used in this invention's algorithm is a passive-mode motor-driven lower limb exoskeleton robot, such as... Figure 3As shown, the system includes a waist structure, a thigh section, a lower leg section, a hip joint mechanism, a knee joint mechanism, and a drive motor located within the joint mechanism for driving joint rotation. A controller is located on a host computer and executes the trajectory tracking control method for a lower limb rehabilitation exoskeleton robot based on super-twisted end sliding mode, as described in the above embodiment.

[0073] The lower limb exoskeleton robot of this invention can complete the designated rehabilitation movement trajectory well, provide a good treatment plan for patients who have completely lost their mobility, and play a role in promoting recovery.

[0074] In summary, this invention proposes a trajectory tracking control method for lower limb rehabilitation exoskeleton robots based on super-twisted end-sliding mode. High-dimensional data such as joint angles and angular velocities are obtained through sensors and input into a designed controller. The super-twisted end-sliding mode algorithm calculates reference torque in real time, ensuring the exoskeleton robot's trajectory conforms to a predetermined path, thereby improving the rehabilitation efficiency of the lower limb rehabilitation exoskeleton robot. Compared to other existing technologies, this method supports high-dimensional input / output, allows for uncertain system parameters, exhibits strong anti-disturbance capabilities, and minimizes control input jitter.

[0075] To more intuitively illustrate the solution and advantages of the present invention, the technical solution of the present invention will be further described below in conjunction with embodiments.

[0076] The following parameters were identified based on the actual lower limb rehabilitation exoskeleton robot.

[0077] symbol describe value <![CDATA[m1]]> thigh bar mass 6.232kg <![CDATA[m2]]> Lower leg weight 6.157kg <![CDATA[l1]]> Thigh bar length 0.43m <![CDATA[l2]]> Lower leg length 0.46m <![CDATA[I1]]> thigh bar rotational inertia <![CDATA[0.095kg·m 2 ]]> <![CDATA[I2]]> Moment of inertia of the lower leg <![CDATA[0.0087kg·m 2 ]]> <![CDATA[l c1 ]]> Thigh bar center of mass distance from joint length 0.18m <![CDATA[l c2 ]]> The distance from the center of mass of the lower leg to the joint length 0.18m

[0078] Thigh gait tracking effect as Figure 4 As shown, the lower leg gait tracking effect is as follows: Figure 5 As shown in the diagram, the tracking error is illustrated below. Figure 6 , Figure 7 As shown, the lower limb rehabilitation exoskeleton robot can achieve good tracking of a given gait trajectory with small tracking error and reduced jitter in the control input, indicating that the method proposed in this invention can achieve good control performance.

[0079] The embodiments described in this specification are merely illustrative examples of implementations of the inventive concept. The scope of protection of this invention should not be considered limited to the specific forms described in these embodiments; rather, it extends to equivalent technical means conceivable by those skilled in the art based on the inventive concept.

Claims

1. A control method for a lower limb rehabilitation exoskeleton robot based on super-twisted end sliding mode, comprising the following steps: (1) Using the motor output torque as the system control input signal, a dynamic model of a single-leg lower limb exoskeleton is established, and it is transformed into a second-order differential equation of the lower limb exoskeleton robot by the Euler-Lagrange method to complete the system modeling. (2) Design a non-singular terminal sliding surface s based on the lower limb rehabilitation exoskeleton robot model; (3) Based on the established dynamic model of the lower limb rehabilitation exoskeleton robot, and the constructed non-singular terminal sliding surface, a super-torsional sliding mode control algorithm is designed; specifically including: (3.1) When designing the sliding mode control law, the equivalent control law is designed based on the designed sliding mode surface. First, it is assumed that the system parameter uncertainty and external disturbance d = [d1d2] T =[0 0] T Combining the non-singular terminal sliding surface s, the equivalent control law is designed as follows: in (3.2) Based on the sliding mode control principle and the super-twist algorithm, the super-twist reaching law is designed as follows: Where k1 and k2 are positive constants; (3.3) Finally, combining the two control laws, a super-twisted sliding mode control algorithm is designed, and the final control input is as follows: τ=τ1+τ2 (14) (3.4) Substituting the obtained control input τ into the Lyapunov function, we get: When the system disturbance d is 0, we can obtain: Based on k1,k2>0, we can conclude that the first derivative of the Lyapunov function is less than zero, and the system state converges in finite time. When the system disturbance d is not 0, then We obtain that in k1-ds -1 / 2 When sgn(s) > 0, the system is stable, and the sliding variable s can converge to the region (dk1) in a finite time. -1 ) 2 Within, that is, the system tracking error e can converge to the region (dk1). -1 ) 2 / γ Inside; (4) Initialize the lower limb exoskeleton robot and determine the exoskeleton's motion mode through a signal input device; the exoskeleton's motion mode is set to walking. (5) Input the signal collected by the joint motor encoder to the super-twisted terminal sliding mode controller. The controller will calculate the motor output torque that has a good tracking effect on the desired gait trajectory based on the information returned by the sensor and according to the current exoskeleton joint angle and the set joint angle.

2. The control method for a lower limb rehabilitation exoskeleton robot based on super-twisted end sliding mode as described in claim 1, characterized in that, Step 1 specifically includes: (1.1) Constructing equations relating joint angles and positions in a lower limb rehabilitation exoskeleton robot: The total kinetic energy and total potential energy of the system are obtained as follows: In the lower limb exoskeleton mechanical leg model, m1 is the mass of the thigh rod, l1 is the length, and l c1 m1 is the distance from the pivot to the center of mass of the thigh; m2 is the mass of the lower leg bar; l2 is the length; l c2 Let θ1 be the distance from the pivot to the center of mass of the lower leg, and θ2 be the angular positions of the hip joint and knee joint, respectively; let p1(x1,y1) and p2(x2,y2) be the coordinates of the center of mass of the thigh and lower leg, respectively. (1.2) Dynamic modeling is performed using the Lagrange method; the Lagrange variables are constructed as follows: L=E k -AND p (3) Substituting this into the following partial differential equation: The Lagrange equations for the thigh and calf are obtained as follows: (1.3) The standard form of the second-order differential equation of the lower limb exoskeleton robot dynamic model is as follows: Where p represents the position vector of the exoskeleton joint, d is the set of external perturbations, and M... p C p and G p Represent the inertia matrix, Coriolis force and centrifugal force matrices, and gravitational torque vector, respectively; in addition, M p C p and G p Specifically, it is expressed as follows: Where τ is the control input torque that needs to be designed in the second step.

3. The control method for a lower limb rehabilitation exoskeleton robot based on super-twisted end sliding mode as described in claim 1, characterized in that, Step 2 specifically includes: (2.1) Define the error variable: (2.2) Taking the derivative with respect to the error variable, we obtain the expression for the first derivative of the error variable: (2.3) Based on the above technical solution, the non-singular fast terminal sliding surface is constructed as follows: Based on the designed non-singular fast terminal sliding surface, the system state moves onto the sliding surface s. When the system state moves onto and remains on the sliding surface s, the system tracking error e converges to zero. The power definition of the column matrix is ​​given as follows: e ρ =[e1 ρ e2 ρ ]; Define Lyapunov functions Taking the first derivative of this function, we get:

Images