Lower limb rehabilitation robot active compliance control method and system thereof

Abstract

The present application relates to a kind of lower limb rehabilitation robot active compliance control method and system thereof.The pressure signal, torque signal, angle signal and surface myoelectric signal are obtained;The estimated value of stiffness coefficient is obtained according to torque signal and angle signal, which is used to control the rigidity degree of feedback system;The estimated value of damping coefficient is obtained according to the estimated value of stiffness coefficient, which is used to control the damping degree of feedback system;The compliance control parameter value is obtained according to surface myoelectric signal, which is used to control the compliance degree of feedback system;The angular acceleration reflecting leg movement trend is obtained according to the stiffness coefficient, damping coefficient and compliance control parameter value;The angular acceleration is input into the control module of motor, and the movement direction and speed of motor are adjusted in real time, to control the compliant movement of rehabilitation robot.The present application has the effect of improving the compliance control of lower limb rehabilitation robot.

Description

Technical Field

[0001] This invention belongs to the field of robot control, and in particular relates to an active compliant control method and system for a lower limb rehabilitation robot. Background Technology

[0002] In the field of modern medicine, rehabilitation training plays a vital role in helping patients restore muscle function and motor ability. The application of robotics technology brings both new opportunities and challenges to medical rehabilitation.

[0003] Currently, rehabilitation robots have become a research hotspot in the medical and rehabilitation fields, especially robots specifically designed for lower limb rehabilitation. With continuous technological advancements, rehabilitation robot technology has received widespread attention and research. Several lower limb rehabilitation robots, such as ReWalk and Ekso, are now available on the market. These products generally employ passive control methods, primarily relying on external assistance to help users walk.

[0004] Regarding the aforementioned technologies, traditional rehabilitation robots typically employ a periodic, fixed control strategy. This control method cannot be adjusted in real time according to the user's actual condition and lacks a human-machine mutual adaptation and human-machine collaborative control mechanism, resulting in unsatisfactory rehabilitation outcomes. Summary of the Invention

[0005] The purpose of this invention is to improve the compliant control effect of lower limb rehabilitation robots, and to provide an active compliant control method for lower limb rehabilitation robots, using the following technical solution:

[0006] In a first aspect, the present invention provides an active compliant control method for a lower limb rehabilitation robot, comprising:

[0007] S1: Acquire preset device pressure signal, torque signal, angle signal and surface electromyography signal;

[0008] S2: Input the torque signal and angle signal into the adaptive stiffness coefficient estimation model to obtain the estimated value of the stiffness coefficient;

[0009] S3: Input the stiffness coefficient estimate into the adaptive damping coefficient estimation model to obtain the damping coefficient estimate;

[0010] S4: Input the pressure signal and surface electromyography signal into the compliance coefficient estimation model to obtain the compliance control parameter value, which is used to provide feedback on the compliance degree of the system control;

[0011] S5: Input the obtained stiffness coefficient, damping coefficient, and compliance control parameter values ​​into the angular acceleration estimation model to obtain the angular acceleration estimate that reflects the leg movement trend;

[0012] S6: Input the estimated angular acceleration value into the motor control module to adjust the motor's direction and speed in real time to control the rehabilitation robot's compliant movement.

[0013] Furthermore, the muscle groups for which surface electromyography signals are collected in step S1 are the rectus femoris, vastus lateralis, and vastus medialis.

[0014] Preferably, the sampling rate of the surface electromyography signal is 1000 Hz.

[0015] Further, in step S2, the torque signal and angle signal are input into the adaptive stiffness coefficient estimation model to obtain the estimated value of the stiffness coefficient, specifically as follows:

[0016] A PD controller is configured to realize the flexion and extension movements of the lower limb rehabilitation robot. The controller controls the desired angle and angular velocity to obtain appropriate proportional and differential terms.

[0017] The robot's current angular velocity is approximated by dividing the difference between the robot angle signal acquired in the previous time step and the robot angle signal in the current time step by the time step length.

[0018] Calculate the angular velocity error and angle error based on the angular velocity signal and angle signal;

[0019] Based on the characteristics of the robot, the position Δx of the robot's center of mass during its motion is obtained as follows:

[0020] Δx=lsin(θ) (1)

[0021] Where l represents the length of the robot's support surface, and θ represents the robot's joint angle;

[0022] The robot's inertia coefficient I is obtained based on the position of its center of mass during its motion, specifically:

[0023] I = Δx 2 m (2)

[0024] Where m represents the mass of the robot;

[0025] The obtained inertia coefficient, torque signal, and angle signal are input into the adaptive stiffness estimation model to obtain the stiffness coefficient estimate K.

[0026]

[0027] Among them, K p K is the proportional term. d Let I be the differential term, T be the torque signal, Δx be the position of the robot's center of mass during motion, and e be the differential term. pe represents the difference between the expected angle and the actual angle. p =θ d -θ, e v This represents the difference between the expected angular velocity and the actual angular velocity.

[0028] Furthermore, the method for inputting the stiffness coefficient estimate into the adaptive damping coefficient estimation model in step S3 to obtain the damping coefficient estimate is as follows:

[0029] The obtained inertia coefficient I and stiffness coefficient estimate K are input into the damping estimation model to obtain the damping coefficient estimate D, specifically:

[0030]

[0031] Where ζ represents the damping ratio, and the set range is [0, 1].

[0032] Furthermore, the method for outputting the pressure signal and surface electromyography signal into the compliance coefficient estimation model in step S4 to obtain the compliance control parameters is as follows:

[0033] Bandpass filtering was applied to the surface electromyography (EMG) signal to obtain the preprocessed surface EMG signal.

[0034] The bandpass filtering method is specifically as follows:

[0035]

[0036] Where y(t) is the filtered output signal, X(jω) is the Fourier transform of the surface electromyography signal in the frequency domain, H(jω) is the transfer function of the filter in the frequency domain, and F... -1 The inverse Fourier transform converts the signal from the frequency domain to the time domain, where j is the imaginary unit, ω represents the relationship between the signal in the time and frequency domains, and t is time.

[0037] The preprocessed surface electromyography (EMG) signals were normalized to obtain the maximum value of each muscle after processing during the entire movement cycle without rehabilitation robot assistance. This value was denoted as the maximum excitability S of the muscle measured by the current channel. max j, specifically:

[0038] S max j = maxS i j (6)

[0039] Where j is the selected channel, i is the signal value of the i-th sampling point, and S is the normalized surface electromyography signal.

[0040] The muscle activation level k is calculated based on the ratio of the intensity of the pre-processed surface electromyography signal to the maximum excitation level, specifically:

[0041]

[0042] EMG represents the current surface electromyography signal value after filtering and normalization.

[0043] The degree of muscle activation is input into the compliance coefficient estimation model, and the compliance control parameter value is dynamically adjusted by combining pressure and surface electromyography signals to obtain the compliance control parameter V.

[0044]

[0045] Where α is the weighting factor,

[0046] in This indicates a movement in the direction of lower limb extension. This indicates that the lower limbs maintain their original speed of movement. It indicates moving in the opposite direction of extending the lower limbs.

[0047] Furthermore, in step S5, the obtained stiffness coefficient K, damping coefficient D, and compliance control parameter V are input into the angular acceleration estimation model to obtain the estimated angular acceleration value. Specifically:

[0048]

[0049] in, e is the estimated value of angular acceleration. F The difference between the desired pressure and the measured torque signal is represented by K, D, and V, which represent the estimated stiffness coefficient, estimated damping coefficient, and compliance control parameter value, respectively.

[0050] Furthermore, the angular acceleration estimate obtained in step S6 The angular velocity and direction of the motor are obtained by inputting them into the motor control model. The system outputs the angular velocity and direction and converts them into commands, which are then sent to the lower limb rehabilitation robot to control its active compliance.

[0051] By adopting the above technical solutions, the robot achieves compliant control of the inertia and damping forces generated during movement through stiffness and damping control algorithms. This aims to improve the rehabilitation effect and training efficiency of lower limb rehabilitation robots, reduce the cost of rehabilitation training, and provide more efficient and safer rehabilitation training for professionals and patients in the field of rehabilitation medicine.

[0052] In a second aspect, the present invention provides a motion control device for a lower limb rehabilitation robot, comprising:

[0053] The data acquisition unit acquires pressure signals, torque signals, angle signals, and surface electromyography signals from the lower limb preset device;

[0054] The first calculation unit inputs the torque signal and angle signal into the adaptive stiffness coefficient estimation model to obtain the stiffness coefficient estimate.

[0055] The second calculation unit inputs the stiffness coefficient estimate into the adaptive damping coefficient estimation model to obtain the damping coefficient estimate.

[0056] The third calculation unit inputs the pressure signal and surface electromyography signal into the compliance coefficient estimation model to obtain the compliance control parameter value;

[0057] The fourth calculation unit inputs the obtained stiffness coefficient, damping coefficient, and compliance control parameter values ​​into the angular acceleration estimation model to obtain an estimated angular acceleration value that reflects the leg's motion trend.

[0058] The fifth calculation unit inputs the estimated angular acceleration value into the control module of the lower limb rehabilitation robot motor, adjusts the motor's motion direction and speed in real time, converts it into instructions, and sends them to the lower limb rehabilitation robot to control the robot's compliant movement.

[0059] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform the method described thereon.

[0060] Fourthly, the present invention provides a computing device, including a memory and a processor, wherein the memory stores executable code, and the processor executes the executable code to implement the method described above.

[0061] In summary, the beneficial effects of this invention are as follows:

[0062] 1. Precise control of the robot's dynamic response speed and stability: Based on the stiffness and damping calculation method proposed in this invention, the robot's control force on the patient's lower limb movements can be adjusted in real time, thereby ensuring safety and comfort during the rehabilitation process. This method reduces the risk of injury to the patient during movement while achieving rehabilitation effects.

[0063] 2. This invention proposes the concept of a compliance coefficient, calculated by measuring pressure and surface electromyography signals, which can accurately assess the compliance of a robot. This coefficient better reflects the robot's actual control capability and response speed to the patient's lower limb movements, thereby improving the robot's rehabilitation effect and comfort.

[0064] 3. Improve the robot's motion efficiency: By acquiring the patient's motion intentions, the movement of the motors can be precisely controlled, thereby reducing the robot's energy consumption and the cost of rehabilitation training. Attached Figure Description

[0065] Figure 1This is a schematic diagram of the structure of a lower limb rehabilitation robot according to the present invention;

[0066] Figure 2 This is a flowchart illustrating an active compliant control method for a lower limb rehabilitation robot according to the present invention.

[0067] Figure 3 This is a flowchart of a specific control method for a lower limb rehabilitation robot according to the present invention. Detailed Implementation

[0068] To make the objectives, technical solutions, and advantages of this invention clearer and more intuitive, the invention will be 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 illustrative of the invention and are not intended to limit the invention.

[0069] This invention will adjust the stiffness coefficient and damping coefficient in real time based on the surface electromyography signal acquisition instrument, torque sensor, and angle sensor, in order to provide a more compliant control scheme for the lower limbs to the greatest extent.

[0070] Reference Figure 1 and Figure 2 This invention discloses an active compliant control method for a lower limb rehabilitation robot, comprising the following steps:

[0071] Step S100: Acquire the preset device torque signal, angle signal and surface electromyography signal.

[0072] Torque and angle signals are sensors installed on the lower limb rehabilitation robot. The frequency of the signals can be set to acquire current data in real time. Surface electromyography (EMG) signals are muscle activity potentials collected by an EMG signal acquisition device and can be transmitted to the control system in real time.

[0073] Step S101: Input the torque signal and angle signal into the adaptive stiffness coefficient estimation model to obtain the estimated value of the stiffness coefficient;

[0074] The torque and angle signals are filtered and then input into the stiffness coefficient estimation model, the construction of which is detailed below.

[0075] Step S102: Input the stiffness coefficient estimate into the adaptive damping coefficient estimation model to obtain the damping coefficient estimate.

[0076] The estimated stiffness coefficient is input into the damping coefficient estimation model, the construction of which is detailed below.

[0077] Step S103: Output the surface electromyography signal to the compliance coefficient estimation model to obtain the compliance control parameter value.

[0078] The electromyographic signal is filtered, and the filtered signal is input into the compliance coefficient estimation model. The construction of this model is described in detail below.

[0079] Step S104: Input the obtained stiffness coefficient, damping coefficient, and compliance control parameter values ​​into the angular acceleration estimation model to obtain the angular acceleration. The construction of this model is detailed below.

[0080] Step S105: Input the angular acceleration into the motor control module to adjust the motor's direction and speed in real time.

[0081] Reference Figure 3 A method for estimating the adaptive stiffness coefficient of a lower limb rehabilitation robot includes the following steps:

[0082] Step S200: Configure the PD controller to achieve flexion and extension movements of the lower limb rehabilitation robot. Within the controller, control the desired angle and angular velocity to obtain a suitable proportional term K. p and the differential term K d .

[0083] A model of a rehabilitation robot was built using MATLAB, and a suitable proportional term K was obtained through simulation experiments. p and the differential term K d .

[0084] Step S202: Read the current angle value θ from the angle sensor and calculate the robot's current angular velocity. The angle can be approximated by dividing the difference between the robot's angle value at the previous time step and the robot's angle value at the current time step by the time step size.

[0085]

[0086] Step S203: Calculate the angular velocity error and angle error based on the angular velocity signal and angle signal. Specifically, this is achieved through:

[0087]

[0088] Among them, e θ Indicates angular error, e v This indicates the angular velocity error.

[0089] Step S204: Obtain the position of the robot's center of mass during its motion based on the robot's characteristics, specifically:

[0090] Δx=lsin(θ) (3)

[0091] Where l represents the length of the robot's support surface, and θ represents the joint angle;

[0092] The robot's inertia coefficient is obtained based on the position of its center of mass during its motion, specifically:

[0093] I = Δx 2 m (4)

[0094] Where m represents the mass of the robot;

[0095] Step S205: Obtain the actual torque T between the robot and the support surface by measuring the torque sensor. According to Newton's second law, we can obtain:

[0096]

[0097] Δx represents the distance between the robot's center point and the support surface, K represents the Cartesian stiffness of the interaction, I represents the robot's moment of inertia, l represents the distance from the robot's center point to the joint, and θ... Let θ and angular acceleration represent the robot's angle and angular acceleration, respectively, where:

[0098] Δx=lsin(θ) (6)

[0099] Furthermore, in the above formula Using the expected angle θ d and angular velocity The substitution yields:

[0100] T = IK p e p +IK d e v +KΔxlsin(θ) (7)

[0101] Among them, e p Indicates angular error, e v This indicates the angular velocity error.

[0102] Step S206: Based on the above steps, obtain the torque T and the proportional term K. p Differential term K d I represents the inertia coefficient, θ represents the angle value, and angular velocity represents the angular velocity. Substituting into the formula, we get:

[0103]

[0104] It can obtain the estimated value of stiffness coefficient in real time.

[0105] Reference Figure 2 A method for estimating the damping coefficient of a lower limb rehabilitation robot includes the following steps:

[0106] Step S300: Determine the current angle value θ of the rehabilitation robot.

[0107] Step S301: A single-degree-of-freedom dynamic model of a lower limb rehabilitation robot can be expressed as follows, with the following formula:

[0108]

[0109] Step S302: Calculate the initial damping coefficient D. According to the definition of damping ratio ζ, the initial damping coefficient can be obtained as follows:

[0110]

[0111] Where I is the inertia coefficient, which can be obtained through step S204, and ζ represents the damping ratio, with a set range of [0, 1].

[0112] Reference Figure 3 A method for estimating the compliance control coefficients of a lower limb rehabilitation robot includes the following steps:

[0113] Step S400: Acquire the signal from the surface electromyography signal acquisition device.

[0114] Based on the human body operation model, the muscle groups involved in the rehabilitation process are the rectus femoris, vastus lateralis, and vastus medialis. Signals from these muscle groups were collected at a sampling rate of 1000 Hz.

[0115] Step S401: Bandpass filter the relevant muscle surface electromyography (EMG) signals to obtain the processed relevant muscle surface EMG signals.

[0116] The method for bandpass filtering of surface electromyography signals is as follows:

[0117]

[0118] Where y(t) is the filtered output signal, X(jω) is the Fourier transform of the surface electromyography signal in the frequency domain, H(jω) is the transfer function of the filter in the frequency domain, and F... -1 The inverse Fourier transform converts the signal from the frequency domain to the time domain, where j is the imaginary unit, ω represents the relationship between the signal in the time and frequency domains, and t is time.

[0119] Step S402: Normalize the preprocessed surface electromyography (EMG) signals to obtain the maximum value of each muscle after processing during the entire movement cycle without rehabilitation robot assistance. This value is denoted as the maximum excitability of the muscle measured by the current channel, Smax,j. Specifically:

[0120] S max j = maxS i j (12)

[0121] Where i is the selected channel, i is the signal value of the i-th sampling point, and S is the normalized surface electromyography signal.

[0122] Step S403: Process the filtered current signal and muscle excitation level to obtain the muscle activation level k.

[0123] The degree of muscle activation is simply calculated here as the ratio of the current filtered signal to the degree of muscle excitation, specifically:

[0124]

[0125] Where EMG represents the filtered and normalized current surface electromyography signal value.

[0126] Step S404: Input the degree of muscle activation into the compliance coefficient estimation model to obtain the compliance control parameter value.

[0127] Surface electromyography (EMG) signals reflect the degree of muscle activation, which in turn reflects the muscle's exertion.

[0128] The compliance control parameter values ​​are dynamically adjusted based on the combined force and surface electromyography signals. The adjustment of the compliance parameter values ​​is defined as follows.

[0129]

[0130] Where α is the weighting factor,

[0131] in This indicates a movement in the direction of lower limb extension. This indicates that the lower limbs maintain their original speed of movement. It indicates moving in the opposite direction of extending the lower limbs.

[0132] Reference Figure 3 The method for calculating the desired acceleration includes the following steps:

[0133] Step S500: Based on the interaction force F obtained in step S205, set the magnitude of the desired force and calculate the error e. F .

[0134] Step S501: Obtain the angle error e according to step 203. θ .

[0135] Step S503: Obtain the stiffness coefficient K, damping coefficient D, and compliance control parameter V according to steps 205, 302, and 404.

[0136] Step S504: Input the obtained stiffness coefficient, damping coefficient, and adjustment values ​​of the damping coefficient into the angular acceleration estimation model, specifically as follows:

[0137]

[0138] in, e is the estimated value of angular acceleration. F The difference between the desired pressure and the measured torque signal is represented by K, D, and V, which represent the estimated stiffness coefficient, estimated damping coefficient, and compliance control parameter value, respectively.

[0139] Based on the above formula, the estimated angular acceleration can be obtained.

[0140] Reference Figure 3 The method for controlling the movement of a lower limb rehabilitation robot based on angular acceleration includes the following steps:

[0141] Step S600: Obtain the speed adjustment amount based on the angular acceleration obtained in step 504.

[0142] Specifically, the speed adjustment is calculated based on the time step Δt:

[0143]

[0144] Step S601 generates a command to control the motor based on the speed adjustment amount, changes the speed, and realizes compliant control of the lower limb rehabilitation robot.

[0145] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional modules is used as an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device, and unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0146] This invention provides an active compliant control method for a lower limb rehabilitation robot, specifically achieved through variable parameter impedance control. This avoids the limitation of previous rehabilitation robots that could not provide personalized rehabilitation, and by adjusting parameters in real time during training, it can bring the best rehabilitation effect to the patient.

[0147] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any feature disclosed in this specification (including the abstract and drawings) may be replaced by other equivalent or similar features unless specifically stated otherwise. That is, unless specifically stated otherwise, each feature is only one example of a series of equivalent or similar features.

Claims

1. A motion control device for a lower limb rehabilitation robot, characterized in that... include: The data acquisition unit acquires pressure signals, torque signals, angle signals, and surface electromyography signals from the lower limb preset device; The first calculation unit inputs the torque and angle signals into the adaptive stiffness coefficient estimation model to obtain the stiffness coefficient estimates; specifically: A PD controller is set up to realize the flexion and extension movements of the lower limb rehabilitation robot; the controller is used to control the desired angle and angular velocity to obtain appropriate proportional and differential terms; The robot's current angular velocity is approximated by dividing the difference between the robot angle signal acquired in the previous time step and the robot angle signal in the current time step by the time step length. Based on angular velocity Calculate the angle error e using the angle signal p and angular velocity error e v ; Based on the characteristics of the robot, the position Δx of the robot's center of mass during its motion is obtained as follows: Δx=lsin(θ) (1) Where l represents the length of the robot's support surface, and θ represents the robot's joint angle; The robot's inertia coefficient I is obtained based on the position of its center of mass during its motion, specifically: I=Δx 2 m (2) Where m represents the mass of the robot; The obtained inertia coefficient, torque signal, and angle signal are input into the adaptive stiffness estimation model to obtain the stiffness coefficient estimate K. Among them, K p K is the proportional term. d Let I be the differential term, T be the torque signal, Δx be the position of the robot's center of mass during motion, and e be the differential term. p Represents the desired angle θ d The difference from the actual angle, e p =θ d -θ, e v Represents the desired angular velocity The difference between the actual angular velocity and the actual angular velocity The second calculation unit inputs the stiffness coefficient estimate into the adaptive damping coefficient estimation model to obtain the damping coefficient estimate. The third calculation unit inputs the pressure signal and surface electromyography signal into the compliance coefficient estimation model to obtain the compliance control parameter values; specifically: Bandpass filtering was applied to the surface electromyography (EMG) signal to obtain the preprocessed surface EMG signal. The bandpass filtering method is specifically as follows: Where y(t) is the filtered output signal, X(jω) is the Fourier transform of the surface electromyography signal in the frequency domain, H(jω) is the transfer function of the filter in the frequency domain, and F... -1 The inverse Fourier transform converts the signal from the frequency domain to the time domain, where j is the imaginary unit, ω is the relationship between the signal in the time domain and the frequency domain, and t is time. The preprocessed surface electromyography (EMG) signals were normalized to obtain the maximum value of each muscle after processing during the entire movement cycle without rehabilitation robot assistance. This value was denoted as the maximum excitability S of the muscle measured by the current channel. maxj Specifically: S maxj =maxS i j (6) Where j is the selected channel, i is the signal value of the i-th sampling point, and S is the normalized surface electromyography signal; The muscle activation level k is calculated based on the ratio of the intensity of the pre-processed surface electromyography signal to the maximum excitation level, specifically: Where EMG is the current surface electromyography signal value after filtering and normalization; The degree of muscle activation is input into the compliance coefficient estimation model, and the compliance control parameter value is dynamically adjusted by combining pressure and surface electromyography signals to obtain the compliance control parameter V. Where α is the weighting factor, in This indicates a movement in the direction of lower limb extension. This indicates that the lower limbs maintain their original speed of movement. It indicates movement in the opposite direction of lower limb extension; The fourth calculation unit inputs the obtained stiffness coefficient, damping coefficient, and compliance control parameter values ​​into the angular acceleration estimation model to obtain an estimated angular acceleration value reflecting the leg's motion trend; specifically: Among them, e F The difference between the desired pressure and the measured torque signal is represented by K, D, and V, which represent the estimated stiffness coefficient, estimated damping coefficient, and compliance control parameter value, respectively. The fifth calculation unit inputs the estimated angular acceleration value into the control module of the lower limb rehabilitation robot motor, adjusts the motor's motion direction and speed in real time, converts it into instructions, and sends them to the lower limb rehabilitation robot to control the robot's compliant movement.

2. The apparatus according to claim 1, characterized in that... The muscle groups from which surface electromyography signals were collected were the rectus femoris, vastus lateralis, and vastus medialis.

3. The apparatus according to claim 1, characterized in that... The sampling rate of the surface electromyography signal is 1000 Hz.

4. The apparatus according to claim 1, characterized in that... The method for inputting the stiffness coefficient estimate into the adaptive damping coefficient estimation model to obtain the damping coefficient estimate is as follows: The obtained inertia coefficient I and stiffness coefficient estimate K are input into the damping estimation model to obtain the damping coefficient estimate D, specifically: Where ζ represents the damping ratio, with a set range of [0,1], and K represents the estimated stiffness coefficient.

Images