Muscle force estimation and control method and device for joint trainer and medium

Abstract

The invention discloses a muscle force estimation and control method and device for a joint trainer and a medium. The muscle force estimation method comprises the following steps that the muscle force of a user is estimated according to current feedback torque of a torque sensor and a servo motor; a user training part muscle force model is established, feedback torque of a torque sensor and feedback torque, joint position and joint speed data of a servo motor in the joint movement process of a user are collected, and parameter identification of the user training part muscle force model is conducted through the least square method based on the torque sensor and the feedback torque of the servo motor. And fusing the user muscle force estimated based on the current feedback torque of the torque sensor and the servo motor to obtain user muscle force data. On the basis of the steps, the control method further comprises the step of controlling the equipment to drive the joints to move according to the selected training mode and the user output condition. The requirements for torque measurement and control in different training modes can be met.

Description

Technical Field

[0001] The present application relates to the technical field of medical robots, and in particular to a muscle force estimation and control method, device and medium for a joint trainer. Background Art

[0002] At present, there are more and more products for joint training on the market. In order to meet the needs of users, it is often necessary to have multiple training modes. The requirements for the calculation and control of user muscle strength in different training modes are also quite different. Taking the knee isokinetic training and evaluation system as an example, in the passive mode, it is necessary to detect the user's muscle strength and determine whether the user has high muscle impedance to take protective actions. The accuracy requirements for torque measurement are relatively high, but the requirements for the range of torque measurement are relatively low. In the isokinetic mode, it is necessary to output adaptive resistance according to the user's output to limit the user's joint speed. At the same time, it is necessary to detect the user's output to evaluate the patient's muscle strength. The range of torque measurement is required to be high, but the accuracy requirements for measurement are relatively low. In addition, due to the differences in the joint mobility of different users, it is also necessary to limit the joint angle during training to avoid user injuries. Therefore, the product needs to achieve control in at least three dimensions: torque, speed, and position.

[0003] Chinese patent CN 108553834A discloses an isokinetic muscle training system and its control algorithm, which uses a resistance strain gauge to detect external force, and controls the servo driver to output torque to drive the user's training part to move by judging the force and direction of movement. Chinese patent CN218979607U discloses a knee joint training device, which uses the difference between the current current of the motor and the preset current to adjust the motor torque. It can be found that the current control method of the joint trainer is relatively simple, and the algorithm is relatively complex. It cannot be compatible with multiple training modes at the same time, and whether it is using strain gauges or calculating the relevant muscle force through motor feedback current, it cannot solve the different requirements for torque acquisition accuracy and range under different training modes. Summary of the invention

[0004] The present application provides a muscle strength estimation and control method, device and medium for a joint trainer, wherein a muscle strength estimation method for a joint trainer can obtain a user muscle strength estimation value based on a torque sensor and a servo motor's current feedback torque to meet the trainer's high torque measurement accuracy requirements in a passive mode and high torque measurement range requirements in a constant speed mode; a control method for a joint trainer can control the trainer according to the obtained user muscle strength estimation value, thereby simplifying the control process and being compatible with different training modes.

[0005] The technical solution of this application is as follows:

[0006] On the one hand, the present application provides a muscle strength estimation method for a joint trainer, comprising the following steps:

[0007] S11: Establish muscle strength model of user training parts;

[0008] S12: Identification data collection: collecting feedback torque of the torque sensor and feedback torque of the servo motor, joint position, and joint speed data when the joint trainer moves back and forth at a low speed within the range of motion of the user's joints;

[0009] S13: using the least square method to identify the parameters of the muscle force model of the user's training part based on the torque sensor and the servo motor feedback torque;

[0010] S14: Obtain the user muscle force estimation value based on the torque sensor and the current feedback torque of the servo motor through the muscle force model after parameter identification, and fuse the two.

[0011] Furthermore, in step S11, the muscle strength model of the user's training part is established as follows:

[0012]

[0013] where τ user is the user's muscle strength, τ sensor1 is the torque sensor feedback torque, τ sensor2 is the servo motor feedback torque, mgl is the product of the user's training part gravity and the lever arm, Respectively represent joint position and velocity, τ c is the Coulomb friction torque, τ offset is the torque offset value;

[0014] Furthermore, step S12 is specifically as follows: calibrate a range of joint motion that a user can easily reach, and use the joint trainer to drive the user's joints to perform low-speed reciprocating motion within the range of joint motion, and record the feedback torque of the torque sensor and the feedback torque of the servo motor, the joint position, and the joint speed.

[0015] Further, in step 13: since the joint trainer performs low-speed passive motion within the user's joint range of motion, it is considered that τ user Close to 0, then:

[0016]

[0017] It can be expressed in linear form as follows:

[0018]

[0019] Among them, X is the design matrix, β is the regression coefficient vector, and ε is the error term vector; the least squares method can be used to calculate the torque τ based on the torque sensor feedback torque sensor1 And the servo motor feedback torque τ sensor2 The model parameters β=[mgl,τ c ,τ offset ].

[0020] Furthermore, in step S14, a switching method is used to fuse data, and the user's muscle strength is estimated based on the torque sensor feedback value and the rated range:

[0021]

[0022] When the torque sensor feedback value is greater than the rated range, the torque sensor can no longer guarantee the linearity of the feedback torque. Therefore, the feedback torque of the servo motor is introduced as the input for user muscle force estimation:

[0023]

[0024] Since the rated range of the torque sensor is much larger than the Coulomb friction torque τ c , so the effect of the reducer friction dead zone on the user's muscle force estimation is ignored.

[0025] Furthermore, in step S14, the data is fused using a weighted average method:

[0026]

[0027] Among them, ω1 and ω2 are the weight coefficients of the torque sensor feedback value and the motor feedback torque value, respectively, and are defined according to the following formula:

[0028] Formula 1:

[0029] ω1=(τ limit1 -τ sensor1 ) / (τ limit1 -τ dead1 )

[0030] ω2=1-ω1

[0031] Formula 2:

[0032] ω2=(τ limit2 -τ sensor2 ) / (τ limit2 -τ dead2 )

[0033] ω1=1-ω2

[0034] where τ limit1 , τ limit2 and τ dead1, τ dead2 They are the upper limit and dead zone of measurement respectively.

[0035] Through the above scheme, the present invention can adapt to the different requirements of different working modes for muscle force estimation accuracy and range. Specifically, by fusing the user muscle strength estimated based on the torque sensor and the current feedback torque of the servo motor, the user muscle force estimation value based on the torque sensor and the current feedback torque of the servo motor is obtained, thereby meeting the high torque measurement accuracy requirements of the trainer in passive mode and the high torque measurement range requirements in constant speed mode.

[0036] In another aspect, the present application provides a control method for a joint trainer, comprising the following steps:

[0037] S1: Obtaining a user's muscle strength estimation value by using the muscle strength estimation method for the joint trainer as described above;

[0038] S2: Control the device to drive joint movement according to the selected training mode and the user's output;

[0039] An admittance model is established, and control data is output based on the user muscle strength obtained in step S1 and the selected joint trainer training mode to control the joint trainer to drive the joint movement.

[0040] Furthermore, step S2 specifically includes the following steps:

[0041] S21: Establish the admittance model:

[0042]

[0043] Change the model to the following:

[0044]

[0045] Points can be obtained:

[0046]

[0047] Where M is the inertia coefficient, B is the damping coefficient, K is the stiffness coefficient, q e , is the expected position, velocity, acceleration, τ passive is the passive torque, Δt is the system scanning period;

[0048] S22: When the joint trainer works in the passive mode, a passive trajectory is generated according to the set target position and maximum speed and input into the PID controller, and the output τ passive And input into the admittance model to generate q e Sending it to the servo driver can control the device to move along the desired trajectory;

[0049] When the joint trainer works in isokinetic mode, τ passive =0, and the user's muscle strength τ user Input into the admittance model to generate q e The device movement can be controlled by sending it to the servo driver.

[0050] Based on the above scheme, the control process is simplified and compatible with different training modes by setting an admittance equation that is applicable to both the passive mode and the constant speed mode of the trainer to generate the control quantity.

[0051] Furthermore, step S2 further includes step S23: setting τ passive The limit range is used to realize the torque protection function, and when the patient is detected to have high muscle impedance, τ user >τ passive , the control equipment gradually stops.

[0052] Furthermore, the method further comprises step S3: e , Set one or more limit intervals.

[0053] Through the above scheme, the user can be protected from three dimensions of joint position, speed and acceleration. On the other hand, the present application provides a control device for a joint trainer, comprising:

[0054] A user muscle strength estimation unit is used to estimate the user's muscle strength based on the torque sensor and the current feedback torque of the servo motor; establish a muscle strength model of the user's training part, collect the feedback torque of the torque sensor and the feedback torque of the servo motor, joint position, and joint speed data during the user's joint movement, use the least squares method to perform parameter identification of the user's training part muscle strength model based on the torque sensor and the feedback torque of the servo motor, and fuse the user's muscle strength estimated based on the torque sensor and the current feedback torque of the servo motor to obtain the user's muscle strength data;

[0055] The control unit is used to establish an admittance model, output control data based on the user's muscle strength obtained and combined with the selected training mode to control the device to drive the joint movement.

[0056] On the other hand, the present application provides a joint trainer, including a servo motor, a reducer, a torque sensor, a synchronous wheel, a synchronous belt, a leg support assembly, a processor and a memory, wherein the memory stores computer instructions, and the processor is configured to call and execute the computer instructions to implement the muscle strength estimation and control method for the joint trainer as described above.

[0057] On the other hand, the present application provides a computer-readable medium, which stores computer instructions. When the computer instructions are called and executed by a computer, the muscle strength estimation and control method for a joint trainer as described above is implemented.

[0058] In summary, the beneficial effects of this application are:

[0059] 1. The muscle force estimation method for the joint trainer provided by the present invention can adapt to the different requirements of different working modes for muscle force estimation accuracy and range. Specifically, by fusing the user muscle force estimated based on the torque sensor and the current feedback torque of the servo motor, the user muscle force estimation value based on the torque sensor and the current feedback torque of the servo motor is obtained, which meets the high torque measurement accuracy requirements of the trainer in the passive mode and the high torque measurement range requirements in the constant speed mode;

[0060] 2. The control method for the joint trainer provided by the present invention generates control quantities by setting an admittance equation applicable to both the passive mode and the constant speed mode of the trainer, thereby simplifying the control process and being compatible with different training modes;

[0061] 3. The control method for the joint training device provided by the present invention is to set e , Set the restriction range to protect users from three dimensions: joint position, speed, and acceleration;

[0062] 4. The control method for the joint trainer provided by the present invention can gradually stop the device when high muscle impedance is detected in the patient, which can reflect a certain degree of flexible control. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 It is a schematic diagram of the knee isokinetic training and assessment system;

[0064] Figure 2 It is a flowchart of user muscle strength estimation in this application. DETAILED DESCRIPTION

[0065] The specific implementation of the present application is described in detail below with reference to the accompanying drawings.

[0066] In the following specific implementation manner, a typical joint training device - a knee joint isokinetic training and evaluation system is used as an example for description. Figure 1 As shown, it includes a servo motor 4, a reducer 3, a torque sensor 2, a driving wheel 1, a driven wheel 6, a synchronous belt 5 and a leg support assembly 7. It can be understood that the method provided by the present invention is also applicable to other types of joint training devices / systems.

[0067] A specific embodiment of the present application provides a muscle strength estimation method for a joint trainer, comprising the following steps:

[0068] S11: Establish muscle strength model of user training parts;

[0069] The established user training part muscle strength model is:

[0070]

[0071] where τ user is the user's muscle strength, τ sensor1 is the torque sensor feedback torque, τ sensor2 is the feedback torque of the servo motor, mgl is the product of the gravity of the user's training part and the lever arm. Taking the knee joint isokinetic training and evaluation system as an example, mgl is the user's calf gravity, Respectively represent joint position and velocity, τ c is the Coulomb friction torque, τ offset is the torque offset value.

[0072] S12: Identification data collection: Collect the feedback torque of the torque sensor and the feedback torque of the servo motor, joint position, and joint speed data when the joint trainer performs low-speed reciprocating motion within the user's joint range of motion; specifically: taking the knee joint isokinetic training and evaluation system as an example, calibrate a joint range that the user can easily reach, and use the joint trainer to drive the user's legs to perform low-speed reciprocating motion within the joint range of motion, and record the feedback torque of the torque sensor and the feedback torque of the servo motor, joint position, and joint speed.

[0073] S13: using the least square method to identify the parameters of the muscle force model of the user's training part based on the torque sensor and the servo motor feedback torque;

[0074] Since the joint trainer performs low-speed passive motion within the user's joint range of motion, it is considered that τ user Close to 0, then:

[0075]

[0076] It can be expressed in linear form as follows:

[0077]

[0078] Among them, X is the design matrix, β is the regression coefficient vector, and ε is the error term vector; the least squares method can be used to calculate the torque τ based on the torque sensor feedback torque sensor1 And the servo motor feedback torque τ sensor2 The model parameters β=[mgl,τ c ,τ offset ].

[0079] In this example, typical results are as follows:

[0080] β1=[7.60,0.4452,0.25]

[0081] β2=[7.24,41.86,0.56]

[0082] It can be seen that the mgl identified by the two are not much different, but because the motor feedback torque includes the reducer friction, τ c There is a big difference.

[0083] S14: Obtain the user muscle force estimation value based on the torque sensor and the current feedback torque of the servo motor through the muscle force model after parameter identification, and fuse the two.

[0084] The switching method is used to fuse the data and estimate the user's muscle force based on the torque sensor feedback value and the rated range:

[0085]

[0086] When the torque sensor feedback value is greater than the rated range, the torque sensor can no longer guarantee the linearity of the feedback torque. Therefore, the feedback torque of the servo motor is introduced as the input for user muscle force estimation:

[0087]

[0088] Since the rated range of the torque sensor is much larger than the Coulomb friction torque τ c , so the effect of the reducer friction dead zone on the user's muscle force estimation is ignored.

[0089] The muscle force estimation method for the joint trainer of this embodiment can adapt to the different requirements of different working modes for muscle force estimation accuracy and range. Specifically, by fusing the user muscle force estimated based on the torque sensor and the current feedback torque of the servo motor, the user muscle force estimation value based on the torque sensor and the current feedback torque of the servo motor is obtained, which meets the high torque measurement accuracy requirements of the trainer in the passive mode and the high torque measurement range requirements in the constant speed mode.

[0090] In some other embodiments, in step S14, the data are fused using a weighted average method:

[0091]

[0092] Among them, ω1 and ω2 are the weight coefficients of the torque sensor feedback value and the motor feedback torque value, respectively, and are defined according to the following formula:

[0093] Formula 1:

[0094] ω1=(τ limit1 -τ sensor1 ) / (τ limit1 -τ dead1 )

[0095] ω2=1-ω1

[0096] Formula 2:

[0097] ω2=(τ limit2 -τ sensor2 ) / (τ limit2 -τ dead2 )

[0098] ω1=1-ω2

[0099] where τ limit1 , τ limit2 and τ dead1 , τ dead2 They are the upper limit and dead zone of measurement respectively.

[0100] Another specific embodiment of the present application provides a control method for a joint trainer, comprising the following steps:

[0101] S1: Obtaining a user's muscle strength estimation value by using the muscle strength estimation method for the joint trainer as described in the above embodiment;

[0102] S2: Control the device to drive joint movement according to the selected training mode and the user's output;

[0103] An admittance model is established, and control data is output based on the user muscle strength obtained in step S1 and the selected joint trainer training mode to control the joint trainer to drive the joint movement.

[0104] Step S2 specifically includes the following steps:

[0105] S21: Establish the admittance model:

[0106]

[0107] Change the model to the following:

[0108]

[0109] Points can be obtained:

[0110]

[0111] Where M is the inertia coefficient, B is the damping coefficient, K is the stiffness coefficient, q e , is the expected position, velocity, acceleration, τ passiveis the passive torque, Δt is the system scanning period;

[0112] S22: When the joint trainer works in the passive mode, a passive trajectory is generated according to the set target position and maximum speed and input into the PID controller, and the output τ passive And input into the admittance model to generate q e Sending it to the servo driver can control the device to move along the desired trajectory;

[0113] When the joint trainer works in isokinetic mode, τ passive =0, and the user's muscle strength τ user Input into the admittance model to generate q e The device movement can be controlled by sending it to the servo driver.

[0114] Based on the above scheme, the control process is simplified and compatible with different training modes by setting an admittance equation that is applicable to both the passive mode and the constant speed mode of the trainer to generate the control quantity.

[0115] Step S23: Setting τ passive The limit range is used to realize the torque protection function, and when the patient is detected to have high muscle impedance, τ user >τ passive , the control equipment gradually stops.

[0116] S3: For q e , Set a limit interval for one or more of the output values ​​in order to e , The output value is limited to the set limit range. e , When the input value exceeds the limit range, q e , The output value is still within the limit range, and the user can be protected from three dimensions: joint position, speed, and acceleration.

[0117] Another specific embodiment of the present application provides a control device for a joint trainer, comprising:

[0118] A user muscle strength estimation unit is used to estimate the user's muscle strength based on the torque sensor and the current feedback torque of the servo motor; establish a muscle strength model of the user's training part, collect the feedback torque of the torque sensor and the feedback torque of the servo motor, joint position, and joint speed data during the user's joint movement, use the least squares method to perform parameter identification of the user's training part muscle strength model based on the torque sensor and the feedback torque of the servo motor, and fuse the user's muscle strength estimated based on the torque sensor and the current feedback torque of the servo motor to obtain the user's muscle strength data;

[0119] The control unit is used to establish an admittance model, output control data based on the user's muscle strength obtained and combined with the selected training mode to control the device to drive the joint movement.

[0120] Another specific embodiment of the present application provides a joint trainer, including a servo motor, a reducer, a torque sensor, a synchronous wheel, a synchronous belt, a leg support assembly, a processor and a memory, wherein the memory stores computer instructions, and the processor is configured to call and execute the computer instructions to implement the muscle strength estimation and control method for the joint trainer as described above.

[0121] Another specific embodiment of the present application provides a computer-readable medium, which stores computer instructions. When the computer instructions are called and executed by a computer, the muscle strength estimation and control method for the joint trainer as described above is implemented.

[0122] The above is only a preferred implementation of the present application. It should be pointed out that a person skilled in the art can make several modifications and improvements without departing from the inventive concept of the present application, and these all fall within the scope of protection of the present application.

Claims

1. A muscle strength estimation method for a joint trainer, characterized in that: The following steps are involved: S11: Establish muscle strength model of user training parts; S12: Identification data collection: collecting feedback torque of the torque sensor and feedback torque of the servo motor, joint position, and joint speed data when the joint trainer moves back and forth at a low speed within the range of motion of the user's joints; S13: using the least square method to identify the parameters of the muscle force model of the user's training part based on the torque sensor and the servo motor feedback torque; S14: Obtain the user muscle force estimation value based on the torque sensor and the current feedback torque of the servo motor through the muscle force model after parameter identification, and fuse the two.

2. The muscle strength estimation method for a joint trainer according to claim 1, characterized in that: In step S11, the muscle strength model of the user's training part is established as: where τ user is the user's muscle strength, τ sensor1 is the torque sensor feedback torque, τ sensor2 is the servo motor feedback torque, mgl is the product of the user's training part gravity and the lever arm, q, Respectively represent joint position and velocity, τ c is the Coulomb friction torque, τ offset is the torque offset value.

3. The muscle strength estimation method for a joint trainer according to claim 1, characterized in that: Step S12 specifically includes: calibrating a joint motion range that a user can easily reach, and using the joint trainer to drive the user's joint to perform low-speed reciprocating motion within the joint motion range, recording the feedback torque of the torque sensor and the feedback torque of the servo motor, the joint position, and the joint speed.

4. The muscle strength estimation method for a joint trainer according to claim 2, characterized in that: In step 13: Since the joint trainer performs low-speed passive motion within the user's joint range of motion, it is considered that τ user Close to 0, then: It can be expressed in linear form as follows: Among them, X is the design matrix, β is the regression coefficient vector, and ε is the error term vector. The least squares method can be used to calculate the torque τ based on the torque sensor feedback torque. sensor1 And the servo motor feedback torque τ sensor2 The model parameters β = [mgl, τ c , τ offset ].

5. The muscle strength estimation method for a joint trainer according to claim 3, characterized in that: In step S14, a switching method is used to fuse data, and the user's muscle strength is estimated based on the torque sensor feedback value and the rated range: When the torque sensor feedback value is greater than the rated range, the torque sensor can no longer guarantee the linearity of the feedback torque. Therefore, the feedback torque of the servo motor is introduced as the input for user muscle force estimation: Since the rated range of the torque sensor is much larger than the Coulomb friction torque τ c , so the effect of the reducer friction dead zone on the user's muscle force estimation is ignored.

6. The muscle strength estimation method for a joint trainer according to claim 3, characterized in that: In step S14, the data is fused using a weighted average method: Among them, ω1 and ω2 are the weight coefficients of the torque sensor feedback value and the motor feedback torque value, respectively, and are defined according to the following formula: Formula 1: ω1=(τ limit1 -t sensor1 ) / (t limit1 -t dead1 ) ω2=1-ω1 Formula 2: ω2=(τ limit2 -t sensor2 ) / (t limit2 -t dead2 ) ω1=1-ω2 where τ limit1 , τ limit2 and τ dead1 , τ dead2 They are the upper limit and dead zone of measurement respectively.

7. A control method for a joint training device, characterized in that: The following steps are involved: S1: Obtaining a user's muscle strength estimation value by using the muscle strength estimation method for a joint trainer according to any one of claims 1 to 6; S2: Control the device to drive joint movement according to the selected training mode and user output; An admittance model is established, and control data is output based on the user muscle strength obtained in step S1 and the selected joint trainer training mode to control the joint trainer to drive the joint movement.

8. The control method for the joint training device according to claim 7, characterized in that: Step S2 specifically includes the following steps: S21: Establish the admittance model: Change the model to the following: Points can be obtained: Where M is the inertia coefficient, B is the damping coefficient, K is the stiffness coefficient, q e , is the expected position, velocity, acceleration, τ passive is the passive torque, Δt is the system scanning period; S22: When the joint trainer works in passive mode, a passive trajectory is generated according to the set target position and maximum speed and input into the PID controller, outputting τ passive And input into the admittance model to generate q e Sending it to the servo driver can control the device to move along the desired trajectory; When the joint trainer works in isokinetic mode, τ passive =0, and the user's muscle strength τ user Input into the admittance model to generate q e The device movement can be controlled by sending it to the servo driver.

9. The control method for the joint training device according to claim 8, characterized in that: Step S2 also includes step S23: setting τ passive The limit range is used to realize the torque protection function, and when the patient is detected to have high muscle impedance, τ user >τ passive , the control equipment gradually stops.

10. The control method for the joint training device according to claim 8, characterized in that: The step S3 is also included: e , Set one or more limit intervals.

11. A control device for a joint trainer, characterized in that: include: A user muscle force estimation unit, used to estimate the user muscle force according to the torque sensor and the current feedback torque of the servo motor; Establish a muscle strength model of the user's training part, collect the feedback torque of the torque sensor and the feedback torque of the servo motor, joint position, and joint speed data during the user's joint movement, use the least squares method to identify the parameters of the muscle strength model of the user's training part based on the torque sensor and the feedback torque of the servo motor, and fuse the user's muscle strength estimated based on the torque sensor and the current feedback torque of the servo motor to obtain the user's muscle strength data; The control unit is used to establish an admittance model, output control data based on the user's muscle strength obtained and combined with the selected training mode to control the device to drive the joint movement.

12. A joint training device, characterized in that: It includes a servo motor, a reducer, a torque sensor, a synchronous wheel, a synchronous belt, a leg support assembly, a processor and a memory, wherein the memory stores computer instructions, and the processor is configured to call and execute the computer instructions to implement the control method for the joint trainer as described in any one of claims 7-10.

13. A computer readable medium, characterized in that The computer-readable medium stores computer instructions, and when the computer instructions are called and executed by a computer, the control method for the joint trainer as described in any one of claims 7 to 10 is implemented.

Images