A knee joint exoskeleton anthropomorphic balance power-assistance control method and device

Abstract

The present invention belongs to the technical field related to wearable exoskeletons, and discloses a knee joint exoskeleton anthropomorphic balance assistance control method and device, which comprises the following steps: (1) calculating the deviation of the speed and position of the human body center of mass and the human body balance reference point, and then obtaining the relative offset speed and displacement of the human body center of mass; (2) obtaining the human body knee joint torque prediction value based on the relative offset speed and displacement of the human body center of mass and the knee joint torque prediction model; (3) dynamically setting the auxiliary gain to dynamically control the magnitude of the auxiliary torque; (4) determining the output auxiliary torque according to the current relative offset of the human body center of mass and the auxiliary torque; (5) converting the auxiliary torque into the corresponding motor current expected value, and then tracking and controlling the input motor current according to the obtained current error, so as to realize real-time control of the motor. The present invention solves the problem that the existing balance assistance method is not anthropomorphic and cannot meet the demand for real-time response.

Description

Technical Field

[0001] The present invention belongs to the technical field related to wearable exoskeletons, and more specifically, relates to a method and device for controlling anthropomorphic balance assistance of a knee exoskeleton. Background Art

[0002] With the advent of an aging society, the global demand for assistive devices is rapidly increasing. In fields such as medicine, the military, and industry, there is an urgent need for technologies and equipment that can provide balance assistance. Especially in the elderly, due to the deterioration of physical functions, balance ability is significantly reduced, which can easily lead to falls and related secondary injuries. Furthermore, the working environment in the military and industrial sectors is often challenging, placing a heavy physical strain on practitioners. Balance-assistance devices are of great significance in reducing their workload and improving work efficiency.

[0003] Currently, existing knee joint torque prediction methods are primarily based on inverted pendulum models. When generating assistive torque, these models often exhibit significant discrepancies from the torque variation patterns of natural human motion, resulting in suboptimal assistive effects and low user comfort. Furthermore, while some spline-based assistive models based on human kinematic models can provide torques that closely mimic natural human motion, these models typically require complex computations and sensor systems for input, which can easily introduce system latency and fail to meet real-time response requirements. Therefore, developing a real-time, anthropomorphic balance-assisted knee exoskeleton control method has become both a key and challenging area of ​​current research. Summary of the Invention

[0004] In response to the above defects or improvement needs of the existing technology, the present invention provides a knee exoskeleton anthropomorphic balance assistance control method and device, which aims to solve the problems that the existing balance assistance methods are not anthropomorphic and cannot meet the needs of real-time response.

[0005] To achieve the above object, according to one aspect of the present invention, a method for controlling anthropomorphic balance of a knee joint exoskeleton is provided, the method comprising the following steps:

[0006] (1) Based on the kinematic data of the human body center of mass and the human body balance reference point in the global coordinate system, the deviation of the velocity and position of the human body center of mass and the human body balance reference point is calculated, and then the relative offset velocity and displacement of the human body center of mass are obtained;

[0007] (2) Inputting the relative offset velocity and displacement of the human body's center of mass into a knee joint torque prediction model, the knee joint torque prediction model outputs a predicted value of the human body's knee joint torque; wherein the mathematical expression of the knee joint torque prediction model is:

[0008]

[0009] Where, is the current moment t after the human nerve delay τ T The predicted value of the human knee joint torque after time, ΔCoM(t) and They are the relative displacement of the human body mass center and the relative offset speed of the human body mass center at the current time t, K p and K v It is the fitting proportional coefficient obtained by performing a binary linear regression on the change in the human knee joint torque in the human balance recovery data and the change in the human center of mass position and velocity;

[0010] (3) The auxiliary gain is dynamically set according to the acceleration detection value of the human body's center of mass at the beginning of the disturbance to dynamically control the magnitude of the auxiliary torque; at the same time, the application time of the auxiliary torque is delayed by designing a delay queue according to the human neural delay time;

[0011] (4) Determine whether the current relative offset of the human body's center of mass exceeds a predetermined severe imbalance threshold. If so, no auxiliary torque is provided, and the output auxiliary torque is zero; otherwise, determine whether the auxiliary torque exceeds a preset maximum auxiliary torque. If so, replace the auxiliary torque at that moment with the maximum auxiliary torque and output it; otherwise, directly output the auxiliary torque at that moment;

[0012] (5) According to the motor and reducer characteristics of the knee exoskeleton, the auxiliary torque is converted into the corresponding expected motor current value, and the current error is obtained by subtracting the expected motor current from the actual motor current. Then, the current of the input motor is tracked and controlled based on the current error to achieve real-time control of the motor.

[0013] Furthermore, the human balance recovery data are the human knee joint torque curve and the human center of mass kinematic curve collected through the rear support surface translation disturbance experiment with different disturbance sizes.

[0014] Furthermore, the assist torque is obtained by multiplying the predicted value of the human knee joint torque by the assist gain; the calculation formula of the assist torque is:

[0015]

[0016] in is the auxiliary torque after dynamic gain adjustment, k base is the basic gain, β is the adjustment factor, and a(0) is the relative offset acceleration of the center of mass of the human body at the initial moment of disturbance.

[0017] Furthermore, the kinematic data of the human body's center of mass in the global coordinate system is the speed and position of the human body's center of mass in the global coordinate system obtained by processing the data measured by the IMU worn on the sacrum; the kinematic data of the human body's balance reference point in the global coordinate system is the speed and position of the human body's balance reference point in the global coordinate system obtained by processing the data measured by the IMU worn on the toes.

[0018] Furthermore, the attitude quaternion of the IMU relative to the global coordinate system is determined, and the data measured by the IMU is converted to the same global coordinate system through attitude quaternion multiplication, thereby obtaining the acceleration values ​​of the human body center of mass and the human body balance reference point, and then performing two integrations to obtain the velocity and displacement of the human body center of mass and the human body balance reference point in the global coordinate system.

[0019] Furthermore, in the initialization phase, the IMU is aligned with the global coordinate system and the attitude quaternion is initialized to (1,0,0,0). The acceleration value a in the corresponding local coordinate system is collected by wearing the IMU at the sacrum and toes of the human body. IMU , the calculation formula for converting the acceleration value in the local coordinate system to the global coordinate system is:

[0020]

[0021] Where, Is a quaternion in the form (0, a ′ x ,a ′ y ,a ′ z ), corresponding to the acceleration value a in the global coordinate system global =(a ′ x ,a ′ y ,a ′ z )m / s 2 ; Represents quaternion multiplication; The quaternion form of the acceleration value in the current IMU local coordinate system; q current =(q0,q1,q2,q3) represents the current attitude quaternion, q current The conjugate quaternion of ;

[0022] After subtracting the z-axis gravity acceleration component, the acceleration of the IMU at the sacrum in the global coordinate system is integrated to obtain the estimated values ​​of the velocity and position of the human center of mass, and the acceleration of the IMU at the toes in the global coordinate system is integrated to obtain the estimated values ​​of the velocity and position of the human balance reference point.

[0023] Furthermore, the relative offset velocity and displacement of the human body's center of mass are smoothed and then input into the knee joint torque prediction model. The step of smoothing the relative offset velocity and displacement of the human body's center of mass includes filtering out the high-frequency noise part through a first-order IIR low-pass filter, and then smoothing the signal through a three-point moving average filter.

[0024] Furthermore, the step of delaying the application of the auxiliary torque by designing a delayed queue includes: constructing a queue of corresponding length according to the update frequency of the auxiliary torque and the target delay time, and each time the auxiliary torque is updated, first reading the auxiliary torque with the arrival delay time corresponding to the current moment from the head of the queue, and then inserting the updated auxiliary torque into the tail of the queue.

[0025] The present invention also provides a knee exoskeleton anthropomorphic balance control system, which includes a memory and a processor. The memory stores a computer program, and the processor executes the knee exoskeleton anthropomorphic balance control method described above when executing the computer program.

[0026] The present invention also provides a computer-readable storage medium, which stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to implement the above-mentioned knee exoskeleton anthropomorphic balance control method.

[0027] In general, the above technical solutions conceived by the present invention, compared with the prior art, provide a knee exoskeleton anthropomorphic balance assistance control method and device with the following advantages:

[0028] 1. The present invention constructs a human knee joint torque prediction model, specifically a human knee joint torque prediction model with the relative offset speed and displacement of the human center of mass as input and the output as the knee joint torque after the nerve delay moment. Based on this knee joint torque prediction model, the human knee joint torque prediction value is calculated, and then the auxiliary torque is calculated, which solves the problem of unnatural and non-anthropomorphic auxiliary torque caused by the traditional power-assisting device starting from the joint kinematics. It can accurately predict the human knee joint torque and provide knee joint assistance consistent with the natural change law of the human body; at the same time, the human knee joint torque prediction model predicts the torque at the next moment in real time based on the data at the current moment, meeting the demand for real-time response.

[0029] 2. The present invention achieves accurate prediction of the human knee joint torque by using only two IMUs as sensors, effectively solving the problem of overly complex sensor systems in traditional balance-assisted exoskeleton systems, thereby simplifying the sensor system of the knee exoskeleton.

[0030] 3. Dynamically adjust the gain coefficient of the assistance according to the intensity of the initial disturbance, so that the exoskeleton can adjust the amplitude of the assistance as needed to ensure safety.

[0031] 4. Relevant safety thresholds are set to prevent the exoskeleton from outputting torque or providing excessive torque when it is unable to restore the person's balance.

[0032] 5. The present invention has the characteristics of high real-time performance, good adaptability and excellent human-computer interaction effect, and can be applied to other joints of the lower limbs. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is a schematic diagram of the functional modules of a knee joint exoskeleton anthropomorphic balance control device provided by the present invention;

[0034] Figure 2 is a schematic diagram of an embodiment of the present invention applied to an anthropomorphic balance assisting device for a knee joint exoskeleton;

[0035] Figure 3 This is a flow chart of a knee exoskeleton anthropomorphic balance control method provided by the present invention;

[0036] Figure 4 Schematic diagram of the construction of a human knee joint torque prediction model provided by an embodiment of the present invention;

[0037] Figure 5 This is a diagram showing the implementation results of the anthropomorphic balance assist control method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0039] See also Figure 3 The present invention provides an anthropomorphic balance control method for a knee exoskeleton. The control method estimates the relative offset speed and position of the human center of mass in real time and utilizes the delayed feedback mechanism of the human center of mass to quickly and flexibly predict the human knee joint torque, provide anthropomorphic torque assistance coordinated with the human balance recovery process, reduce the knee joint force of the elderly after being disturbed, and provide balance and stability of the entire human-machine system.

[0040] The control method mainly includes the following steps:

[0041] Step 1: Based on the kinematic data of the human body center of mass and the human body balance reference point in the global coordinate system, the deviation of the speed and position of the human body center of mass and the human body balance reference point is calculated, and then the relative offset speed and displacement of the human body center of mass are obtained.

[0042] The kinematic data of the human body's center of mass in the global coordinate system is obtained by processing the data measured by the IMU worn on the sacrum to obtain the velocity and position of the human body's center of mass in the global coordinate system; the kinematic data of the human body's balance reference point in the global coordinate system is obtained by processing the data measured by the IMU worn on the toes to obtain the velocity and position of the human body's balance reference point in the global coordinate system; the processing includes determining the attitude quaternion of the IMU relative to the global coordinate system, converting the data measured by the IMU to the same global coordinate system through attitude quaternion multiplication, and then obtaining the acceleration values ​​of the human body's center of mass and the human body's balance reference point, and then performing two integrations to obtain the velocity and displacement of the human body's center of mass and the human body's balance reference point in the global coordinate system.

[0043] During the initialization phase, the IMU is aligned with the global coordinate system and the attitude quaternion is initialized to (1,0,0,0). The acceleration value a in the corresponding local coordinate system is collected by wearing the IMU at the sacrum and toes of the human body. IMU , the calculation formula for converting the acceleration value in the local coordinate system to the global coordinate system is:

[0044]

[0045] Where, Is a quaternion in the form (0, a ′ x ,a ′ y ,a ′ z ), corresponding to the acceleration value a in the global coordinate system global =(a ′ x ,a ′ y ,a ′ z )m / s 2 ; Represents quaternion multiplication; The quaternion form of the acceleration value in the current IMU local coordinate system; q current =(q0,q1,q2,q3) represents the current attitude quaternion, q current The conjugate quaternion of .

[0046] After subtracting the z-axis gravity acceleration component, the acceleration integration of the IMU at the sacrum in the global coordinate system can obtain the estimated values ​​of the human center of mass velocity and position, and the acceleration integration of the IMU at the toes in the global coordinate system can obtain the estimated values ​​of the human balance reference point velocity and position.

[0047] In another embodiment, Figure 2 and Figure 1 As shown, before installation, the IMUs were initially calibrated by aligning their local coordinate systems with the spatial coordinate system in the same environment. When wearing the IMUs, they ensured a tight fit with the sacral S1 joint and the toes, respectively. They functioned as the center of mass kinematic estimation unit and the balance reference point kinematic estimation unit, respectively, to estimate the velocity and position of the center of mass and the balance reference point in the global coordinate system during balance recovery. These estimates were subtracted to obtain the relative offset velocity and displacement of the center of mass. After passing through a low-pass filter and a sliding average filter, these values ​​were used to obtain the input to the knee joint torque prediction unit, thereby generating a predicted value for the knee joint torque. The input torque was converted into a desired motor current, and the actual electrode current was obtained and input into a current controller to drive the motor. The desired motor current was calculated based on the characteristics of the motor and reducer; the reducer was a planetary reducer with a reduction ratio of 1:10. The actual motor current was measured using the motor's ammeter. The current controller included a PID controller that decoupled the actual motor current into d-axis and q-axis currents using Clark and Park transforms, and then generated the control variable using the PID controller. Subsequently, these control quantities are converted into three-phase voltage instructions through inverse transformation, and then the inverter is controlled through the PWM module to achieve a fast and accurate response of the motor to the target current, ensuring efficient and stable operation of the motor, thereby acting on the human knee joint and assisting the human body in regulating its own center of mass.

[0048] Step 2: Smoothing the relative offset velocity and displacement of the human body's center of mass.

[0049] The steps of smoothing the relative offset velocity and displacement of the human body's center of mass include filtering out the high-frequency noise part through a first-order IIR low-pass filter, and then smoothing the signal through a three-point moving average filter to reduce short-term fluctuations and make the final signal more stable and smooth.

[0050] Step 3: Input the processed relative offset velocity and displacement of the human body center of mass into the knee joint torque prediction model, and the knee joint torque prediction model outputs the human body knee joint torque prediction value; wherein the mathematical expression of the knee joint torque prediction model is:

[0051]

[0052] Where, is the current moment t after the human nerve delay τ T The predicted value of the human knee joint torque after time, ΔCoM(t) and They are the relative displacement of the human body mass center and the relative offset speed of the human body mass center at the current time t, K p and K v This is the fitting proportional coefficient obtained by performing a linear regression between the change in knee joint torque and the change in center of mass position and velocity in the balance recovery data. The balance recovery data consists of knee joint torque curves and center of mass kinematic curves collected through rearward support surface translation perturbation experiments of varying magnitudes.

[0053] Step 4: Dynamically set the auxiliary gain based on the human body center of mass acceleration detection value at the start of the disturbance to dynamically control the size of the auxiliary torque; at the same time, according to the human nerve delay time, the application time of the auxiliary torque is delayed by designing a delay queue; among them, the auxiliary torque is obtained by multiplying the human knee joint torque prediction value and the auxiliary gain.

[0054] The calculation formula of the auxiliary torque is:

[0055]

[0056] in is the auxiliary torque after dynamic gain adjustment, k base is the basic gain, β is the adjustment factor, and a(0) is the relative offset acceleration of the center of mass of the human body at the initial moment of disturbance.

[0057] The steps of delaying the application of the auxiliary torque by designing a delayed queue include: building a queue of corresponding length according to the update frequency of the auxiliary torque and the target delay time. Each time the auxiliary torque is updated, the auxiliary torque with the arrival delay time corresponding to the current moment is first read from the head of the queue, and then the updated auxiliary torque is inserted into the tail of the queue.

[0058] Step 5: Determine whether the current relative offset of the human body's center of mass exceeds the predetermined severe imbalance threshold. If so, no auxiliary torque is provided and the output auxiliary torque is zero. Otherwise, determine whether the auxiliary torque exceeds the preset maximum auxiliary torque. If so, replace the auxiliary torque at that moment with the maximum auxiliary torque and output it. Otherwise, directly output the auxiliary torque at that moment.

[0059] In this embodiment, by real-time monitoring of the relative displacement of the center of mass of the user of the knee exoskeleton, a severe imbalance threshold and a maximum assisting torque are set to ensure the safety of the exoskeleton assistance. The range of the relative displacement of the human body's center of mass is determined: if the current relative displacement range of the human body's center of mass exceeds the predetermined severe imbalance threshold, no assistance is provided to avoid additional interference. If the current relative displacement of the human body's center of mass does not exceed the severe imbalance threshold, the magnitude of the assisting torque is further evaluated; when evaluating the magnitude of the assisting torque, if the magnitude of the assisting torque exceeds the preset maximum assisting torque, the assisting torque at that moment is replaced with the maximum assisting torque and then output; if the magnitude of the assisting torque does not exceed the preset maximum assisting torque, the assisting torque at that moment is directly output.

[0060] The safety threshold is divided into a severe imbalance threshold and a maximum assist torque threshold. In this embodiment, the severe imbalance threshold is set to 20% of the human body stability margin, that is, 15 cm, and the maximum assist torque threshold is set to ±30 Nm.

[0061] Step 6: Based on the motor and reducer characteristics of the knee exoskeleton, the auxiliary torque is converted into the corresponding expected motor current value, and the current error is obtained by subtracting the expected motor current from the actual motor current. The current of the input motor is then tracked and controlled based on the current error to achieve real-time control of the motor.

[0062] In one embodiment, see Figure 4 and Figure 5 The knee joint torque prediction model assumes that when the center of mass position change and velocity change are zero, the knee joint torque should also be zero. Therefore, the knee joint torque prediction model only focuses on the change amount without considering the bias. The mathematical expression of the knee joint torque prediction model is:

[0063] A binary linear regression analysis was performed on the knee torque prediction model using data from a perturbation experiment involving a posterior displacement of the support surface. The experiment measured the actual relative center of mass displacement, center of mass velocity, and corresponding knee torque during balance recovery in multiple groups of subjects under varying perturbation magnitudes. This data collection aimed to reveal the quantitative relationship between center of mass motion and knee torque, thereby supporting the validity and applicability of the knee torque prediction model.

[0064] The relative displacement and velocity set X of the center of mass of all samples is expressed as:

[0065]

[0066] The set T of knee joint torques of all samples is expressed as:

[0067] T=[T knee,1 (t+τ T )Tknee,2 (t+τ T )…T knee,n (t+τ T )] T .

[0068] The regression coefficient K in the prediction model is expressed as: K = [K p K v ] T .

[0069] The corresponding least squares objective function S(K) is:

[0070] S(K)=(T-XK) T (T-XK).

[0071] The normal equation obtained by the least squares method is:

[0072] X T XK=X T T.

[0073] After solving, the regression coefficient in the knee joint torque prediction model is obtained, thereby realizing the construction of the human knee joint torque prediction model.

[0074] Figure 5 The auxiliary torque provided by the knee exoskeleton to the human body changes synchronously with the real human knee joint measured by external equipment and the change pattern is consistent.

[0075] The present invention also provides a knee joint exoskeleton anthropomorphic balance control device, which comprises:

[0076] Human center of mass kinematics estimation unit: used to measure the velocity and position of the human center of mass (the sacrum S1 joint of the human body) in the global coordinate system in real time during the balance recovery process;

[0077] Human balance reference point kinematics estimation unit: used in conjunction with the human center of mass kinematics estimation unit, it measures the velocity and position of the balance reference point (the point where the human foot is relatively stationary with the support surface) in the global coordinate system during the balance recovery process in real time, providing kinematic input for calculating the offset of the human center of mass relative to the balance reference point;

[0078] Center of mass offset difference unit: used to calculate the speed and position deviation of the human body's center of mass relative to the balance reference point, and obtain the relative offset speed and displacement of the human body's center of mass;

[0079] Smoothing filter unit: used to perform smoothing filtering on the relative offset speed and displacement of the human body's center of mass to reduce noise interference;

[0080] Human knee joint torque prediction unit: used to calculate the predicted value of knee joint torque based on the speed and position deviation of the human center of mass relative to the balance reference point after smoothing.

[0081] Auxiliary gain dynamic control unit: used to dynamically adjust the auxiliary gain according to the relative acceleration of the human body's center of mass at the initial moment of disturbance, and calculate the knee joint auxiliary torque based on the auxiliary gain;

[0082] Auxiliary torque delay unit: used to delay the auxiliary torque according to the set time;

[0083] Safety protection unit: used to determine whether the relative displacement of the human body's center of mass and the auxiliary torque exceed the set safety threshold;

[0084] Expected motor current generating unit: used for generating expected motor current according to the auxiliary torque value;

[0085] Actual motor current acquisition unit: used to obtain actual motor current data of the motor of the power exoskeleton power assist control device;

[0086] Motor current difference unit: used to make a difference between the expected motor current and the actual motor current to obtain the current error;

[0087] Current controller: used to track and control the motor current based on the current error.

[0088] The present invention also provides a knee exoskeleton anthropomorphic balance control system, which includes a memory and a processor. The memory stores a computer program, and the processor executes the knee exoskeleton anthropomorphic balance control method described above when executing the computer program.

[0089] The present invention also provides a computer-readable storage medium, which stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to implement the above-mentioned knee exoskeleton anthropomorphic balance control method.

[0090] It will be easily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A knee joint exoskeleton anthropomorphic balance control method, characterized in that: The control method comprises the following steps: (1) Based on the kinematic data of the human body center of mass and the human body balance reference point in the global coordinate system, the deviation of the velocity and position of the human body center of mass and the human body balance reference point is calculated, and then the relative offset velocity and displacement of the human body center of mass are obtained; (2) Inputting the relative offset velocity and displacement of the human body's center of mass into a knee joint torque prediction model, the knee joint torque prediction model outputs a predicted value of the human body's knee joint torque; wherein the mathematical expression of the knee joint torque prediction model is: Where, is the predicted value of the human knee joint torque at the current moment t after the human nerve delay τT time, ΔCoM(t) and They are the relative displacement of the human body mass center and the relative offset speed of the human body mass center at the current time t, K p and K v It is the fitting proportional coefficient obtained by performing a binary linear regression on the change in the human knee joint torque in the human balance recovery data and the change in the human center of mass position and velocity; (3) The auxiliary gain is dynamically set according to the acceleration detection value of the human body's center of mass at the beginning of the disturbance, and the auxiliary torque is dynamically obtained based on the predicted value of the human knee joint torque and the auxiliary gain; at the same time, according to the human nerve delay time, the application time of the auxiliary torque is delayed by designing a delay queue; (4) Determine whether the current relative offset of the human body's center of mass exceeds a predetermined severe imbalance threshold. If so, no auxiliary torque is provided, and the output auxiliary torque is zero; otherwise, determine whether the auxiliary torque exceeds a preset maximum auxiliary torque. If so, replace the current auxiliary torque with the maximum auxiliary torque and output it; otherwise, directly output the current auxiliary torque; (5) According to the motor and reducer characteristics of the knee exoskeleton, the auxiliary torque is converted into the corresponding expected motor current value, and the current error is obtained by subtracting the expected motor current from the actual motor current. Then, the current of the input motor is tracked and controlled based on the current error to achieve real-time control of the motor.

2. The knee joint exoskeleton anthropomorphic balance control method according to claim 1, characterized in that: The human balance recovery data includes the human knee joint torque curve and the human center of mass kinematic curve collected through the backward support surface translation disturbance experiment with different disturbance sizes.

3. The knee joint exoskeleton anthropomorphic balance control method according to claim 1, characterized in that: The assist torque is obtained by multiplying the predicted value of the human knee joint torque by the assist gain. The calculation formula of the assist torque is: in is the auxiliary torque after dynamic gain adjustment, k base is the basic gain, β is the adjustment factor, and a(0) is the relative offset acceleration of the center of mass of the human body at the initial moment of disturbance.

4. The knee joint exoskeleton anthropomorphic balance control method according to claim 1, wherein: The kinematic data of the human body's center of mass in the global coordinate system is the speed and position of the human body's center of mass in the global coordinate system obtained by processing the data measured by the IMU worn on the sacrum; the kinematic data of the human body's balance reference point in the global coordinate system is the speed and position of the human body's balance reference point in the global coordinate system obtained by processing the data measured by the IMU worn on the toes.

5. The knee joint exoskeleton anthropomorphic balance control method according to claim 4, characterized in that: Determine the attitude quaternion of the IMU relative to the global coordinate system, and convert the data measured by the IMU to the same global coordinate system through attitude quaternion multiplication, and then obtain the acceleration values ​​of the human body center of mass and the human body balance reference point. After two integrations, the velocity and displacement of the human body center of mass and the human body balance reference point in the global coordinate system are obtained.

6. The knee joint exoskeleton anthropomorphic balance control method according to claim 4, characterized in that: During the initialization phase, the IMU is aligned with the global coordinate system and the attitude quaternion is initialized to (1, 0, 0, 0). The acceleration value a in the corresponding local coordinate system is collected by wearing the IMU at the sacrum and toes of the human body. IMU , the calculation formula for converting the acceleration value in the local coordinate system to the global coordinate system is: Where, Is a quaternion of the form (0, a′ x , a′ y , a′ z ), corresponding to the acceleration value a in the global coordinate system global =(a′ x , a′ y , a′ z )m / s 2 ; Represents quaternion multiplication; The quaternion form of the acceleration value in the current IMU local coordinate system; q current =(q0, q1, q2, q3) represents the current attitude quaternion, q c Current's conjugate quaternion; After subtracting the z-axis gravity acceleration component, the acceleration of the IMU at the sacrum in the global coordinate system is integrated to obtain the estimated values ​​of the velocity and position of the human center of mass, and the acceleration of the IMU at the toes in the global coordinate system is integrated to obtain the estimated values ​​of the velocity and position of the human balance reference point.

7. The knee joint exoskeleton anthropomorphic balance control method according to claim 1, characterized in that: The relative offset velocity and displacement of the human body's center of mass are smoothed and then input into the knee joint torque prediction model. The steps of smoothing the relative offset velocity and displacement of the human body's center of mass include filtering out the high-frequency noise part through a first-order IIR low-pass filter, and then smoothing the signal through a three-point moving average filter.

8. The method for controlling the humanoid balance of a knee joint exoskeleton according to any one of claims 1 to 7, wherein: The steps of delaying the application of the auxiliary torque by designing a delayed queue include: building a queue of corresponding length according to the update frequency of the auxiliary torque and the target delay time. Each time the auxiliary torque is updated, the auxiliary torque with the arrival delay time corresponding to the current moment is first read from the head of the queue, and then the updated auxiliary torque is inserted into the tail of the queue.

9. A knee joint exoskeleton anthropomorphic balance control system, characterized by: The system includes a memory and a processor, the memory stores a computer program, and the processor executes the knee exoskeleton anthropomorphic balance control method according to any one of claims 1 to 8 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores machine-executable instructions. When the machine-executable instructions are called and executed by the processor, the machine-executable instructions prompt the processor to implement the knee exoskeleton anthropomorphic balance control method according to any one of claims 1 to 8.

Images