Personalized upper limb rehabilitation system based on motion primitives

Abstract

The application discloses a kind of personalized upper limb collaborative rehabilitation systems based on motion primitive, the personalized upper limb collaborative rehabilitation systems include host computer module, path planning module, motion primitive planning module, human kinematics solving module, exoskeleton kinematics solving module, controller module and rehabilitation exoskeleton module, the motion primitive planning module is used to convert discrete point on task space according to patient healthy side movement habit into discrete point on task space of patient habit, and the controller module is driven exoskeleton module by executor to assist patient to carry out movement.The personalized bilateral collaborative rehabilitation system based on motion primitive provided in the application is interacted with patient using host computer module, constitutes a kind of personalized bilateral collaborative rehabilitation system based on motion primitive, realizes personalized training task customization in rehabilitation training process, is conducive to patient to complete rehabilitation process gradually, and then enhance the motor function of hemiplegic upper limb.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of upper limb rehabilitation, in particular to a personalized upper limb cooperative rehabilitation system based on motion primitives. BACKGROUND

[0002] Stroke is a disease with high mortality and high disability rate, and hemiplegic upper limb movement disorder is a common sequelae of stroke, which seriously affects the daily life of patients. High-intensity and high-dose exercise rehabilitation therapy plays a key role in the recovery of hemiplegic upper limbs, but the number of stroke patients is too large, and there is a serious shortage of rehabilitation therapists. Traditional manual physical therapy cannot meet the huge rehabilitation needs. Upper limb rehabilitation robots can effectively alleviate the pressure of rehabilitation therapist shortage, improve the efficiency and effectiveness of rehabilitation training, and have a huge clinical application prospect.

[0003] At present, exercise rehabilitation training is the main means of clinical treatment of stroke hemiplegia, and its principle can be explained by the neural plasticity theory in brain science. The theory points out that internal and external environmental factors can affect the correlation between neurons, so as to change it. In other words, the central nervous system of stroke hemiplegia patients after injury has the possibility of structural and functional changes due to environmental changes. Further, a large number of studies in the field of brain science show that exercise is an important factor affecting neural plasticity, which mainly affects the neural plasticity of the brain from three levels of system macro level, cell level and molecular level. Therefore, a large number of repeated exercise training can provide environmental input for the central nervous system, thereby positively affecting the plasticity of the brain from different levels and making the damaged central nervous system recover and the patient regain effective control of the limbs.

[0004] For exercise rehabilitation, the overall goal is to restore the motor function of the affected limb of the patient through planned rehabilitation training, so that the patient regains the ability to live independently and improves the quality of life. During the exercise rehabilitation treatment process, the patient should try to use normal movement patterns for rehabilitation training to prevent the acquisition of incorrect movement habits, such as excessive use of the trunk for movement, etc. At the same time, the treatment plan should be purposeful and planned, and different movement abilities of patients should correspond to different levels of rehabilitation training, and the recovery of the patient should be re-evaluated during the rehabilitation process, so as to adjust the training type and difficulty according to the evaluation results.

[0005] However, the existing upper limb rehabilitation robots lack personalized training for patients and cannot design the movement habits of patients according to their individuality, thereby affecting the rehabilitation training effect of hemiplegic upper limbs. There is currently a lack of an upper limb cooperative rehabilitation system that can be personalized for patients on the market. SUMMARY

[0006] In order to solve the problems in the prior art, the application provides a personalized upper limb cooperative rehabilitation system based on motion primitives, which can achieve the purpose of personalized treatment in the field of rehabilitation through motion primitives.

[0007] In order to achieve the purpose of the application, the application provides a personalized upper limb cooperative rehabilitation system based on motion primitives, which comprises an upper computer module, a path planning module, a motion primitive planning module, a human body kinematics solving module, an exoskeleton kinematics solving module, a controller module and a rehabilitation exoskeleton module.

[0008] The exoskeleton module serves as an actuator of the system and is used for collecting real-time motion parameters of the healthy side and assisting the hemiplegic limb to complete rehabilitation motion training.

[0009] The upper computer module is used for giving a patient a task target point and interacting with the patient in rehabilitation training by using virtual reality.

[0010] The path planning module is used for interpolating between the target point and the starting point and inputting to the motion primitive planning module, which is used for converting discrete points on the task space into discrete points on the task space according to the motion habit of the healthy side of the patient.

[0011] The human body kinematics solving module converts the discrete points on the task space into discrete points in the joint space of the human body, and then converts them into discrete points in the joint space of the exoskeleton through the exoskeleton kinematics solving module.

[0012] The controller module drives the exoskeleton module to assist the patient to move according to the discrete points in the joint space of the exoskeleton through the actuator.

[0013] Further, the exoskeleton module is used for driving the hemiplegic upper limb of the patient to perform rehabilitation motion training, the healthy upper limb first performs appropriate motion task training through the exoskeleton module, the real-time motion parameters of the healthy side are obtained through the exoskeleton module, and the training data are transmitted to the motion primitive planning module as input parameters.

[0014] Further, the model of the motion primitive planning module is:

[0015]

[0016] Wherein, x is the position of the exoskeleton end actuator in the task space, x0 and g respectively represent the initial position and the target position, v represents the velocity, represents the acceleration, K and D are the elastic parameter and the damping parameter of the system, and f is a function of the correction term to be adjusted.

[0017] Further, the correction term function f of the motion primitive planning module is adjusted by using f(s), and f(s) is:

[0018]

[0019] d(s) is a Gaussian function:

[0020]

[0021] where a i is the width value of the Gaussian function, c i is the center value of the Gaussian function, and w i is an adjustable weight.

[0022] Further, the workflow of the motion primitive planning module is as follows:

[0023] ①First, let the healthy side of the patient perform a series of rehabilitation training tasks with the exoskeleton module, denoted as position vector x, and each time sampling point (t = 0, 1, 2, …, T);

[0024] ②Then, according to the position information, the velocity vector v and the acceleration vector a of the trajectory in the task space are calculated;

[0025] ③Calculate s(t)

[0026]

[0027] where α is the motion duration, and τ is the time scaling factor.

[0028] ④Finally, according to s(t) and the model of the motion primitive planning module, the correction term f(s) is set:

[0029]

[0030] Therefore, the weight w i can be solved using the least squares method, and thus f(s) is used to set f.

[0031] min∑ s (f target -f(s))

[0032] Further, the exoskeleton module collects the motion parameters of the healthy side as shoulder and elbow joint motion parameters; the shoulder and elbow joint motion parameters include shoulder joint adduction and abduction angle θ1, shoulder joint flexion and extension angle θ2, shoulder joint internal and external rotation angle θ3, elbow joint flexion and extension angle θ4, and the spatial pose of the end.

[0033] Further, the path planning module is used to calculate the initial point and the end point of the task issued by the upper computer module, and then the starting position and the final position of each joint of the exoskeleton are calculated through inverse kinematics, and then each joint variable in these two states can be obtained through inverse kinematics operation.

[0034] Further, the working process of the human kinematics solving module comprises:

[0035] A DH model is established according to the human shoulder joint and the elbow joint;

[0036] According to the joint coordinates obtained by the path planning module and the motion primitive module, the end position coordinates of the human body are calculated through forward kinematics.

[0037] Further, the working process of the exoskeleton kinematics solving module comprises:

[0038] A DH model is established according to the morphology of the exoskeleton;

[0039] According to the end position coordinates obtained by the human kinematics solving module, the positions of the joints of the exoskeleton are calculated through inverse kinematics.

[0040] Compared with the prior art, the present application has the following advantages and technical effects:

[0041] The present application proposes a personalized bilateral collaborative rehabilitation system based on motion primitives, aiming to solve the problem of how to make the patient complete the rehabilitation process gradually in the training process in the existing upper limb rehabilitation training process. First, the motion parameter data of the healthy side of the patient is collected through the rehabilitation exoskeleton module, the route planning of the affected side is obtained through the motion primitive planning module, and the route planning has the motion characteristics of the healthy side. Then, the rehabilitation exercise training of the hemiplegic upper limb is assisted through the exoskeleton module, and the upper computer module and the patient are interacted to form a personalized bilateral collaborative rehabilitation system based on motion primitives, which realizes the customization of the personalized training task in the rehabilitation training process, is beneficial to the patient to complete the rehabilitation process gradually, and further enhances the motor function of the hemiplegic upper limb. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 It is a principle schematic diagram of the personalized bilateral collaborative rehabilitation system based on motion primitives in the embodiment of the present application.

[0043] Figure 2 It is a working schematic diagram of the path planning module of the personalized bilateral collaborative rehabilitation system based on motion primitives in the embodiment of the present application.

[0044] Figure 3 It is a schematic diagram of the human kinematics and solving module and the exoskeleton kinematics solving module of the personalized bilateral collaborative rehabilitation system based on motion primitives in the embodiment of the present application. DETAILED DESCRIPTION

[0045] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts are within the protection scope of the present application.

[0046] As shown in Figure 1 The application discloses a personalized upper limb cooperative rehabilitation system based on motion primitives, which comprises an upper computer module, a path planning module, a motion primitive planning module, a human body kinematics solving module, an exoskeleton kinematics solving module, a controller module and a rehabilitation exoskeleton module.

[0047] The rehabilitation exoskeleton module serves as an actuator of the system and is used for collecting real-time motion parameters of a healthy side and assisting a hemiplegic limb to complete rehabilitation motion training.

[0048] The upper computer module is used for giving a patient a task target point and interacting with the patient in rehabilitation training by using virtual reality.

[0049] The path planning module is used for interpolating between the target point and a starting point and inputting to the motion primitive planning module.

[0050] The motion primitive planning module is used for converting discrete points on a task space into discrete points on a task space according to a motion habit of the patient, wherein the motion habit comprises real-time motion parameters and historical motion parameters of the healthy side obtained by the rehabilitation exoskeleton module.

[0051] The human body kinematics solving module is used for converting the discrete points on the task space into discrete points of a human body joint space and then converting discrete points of an exoskeleton joint space by the exoskeleton kinematics solving module, wherein a general robot kinematics analysis method is adopted in the conversion process.

[0052] The controller module drives the rehabilitation exoskeleton module to assist the patient to move according to the discrete points of the exoskeleton joint space by an actuator.

[0053] The rehabilitation exoskeleton module is used for driving a hemiplegic upper limb of the patient to perform rehabilitation motion training, and a healthy upper limb is first trained by the rehabilitation exoskeleton module to perform a proper motion task, real-time motion parameters of the healthy side are obtained by the rehabilitation exoskeleton module, and training data are transmitted to the motion primitive planning module as input parameters.

[0054] In some embodiments of the present application, the model of the motion primitive planning module is as follows:

[0055]

[0056] wherein τ is a time scale factor, x is the position of the exoskeleton end effector in the task space, x0 and g represent the initial position and the target position of the exoskeleton end effector respectively, v represents the velocity, represents the acceleration, K and D are the elastic parameter and the damping parameter of the system, and f is a function of the correction term to be tuned.

[0057] The correction term function f of the motion primitive planning module uses f(s) for tuning, and f(s) is:

[0058]

[0059] d(s) is a Gaussian kernel function:

[0060]

[0061] wherein a i is the width value of the Gaussian kernel function, c i is the center value of the Gaussian kernel function, ω i is the adjustable weight, and s is the abbreviation of s(t) and is the dependent variable with the time t as the independent variable, and i is the serial number.

[0062] The calculation flow of the motion primitive planning module is as follows:

[0063] ①First, let the healthy side of the patient execute a series of rehabilitation training tasks with the rehabilitation exoskeleton module, denoted as the position vector x, and each time sampling point (t=0, 1, 2……T), T is the serial number of the discrete time points;

[0064] ②Then, the velocity vector v and the acceleration vector a of the trajectory in the task space are calculated according to the position information;

[0065] ③Calculate s(t)

[0066]

[0067] wherein α is the motion duration, τ is the time scale factor, and s(t) is the unit nonlinear number.

[0068] ④Finally, the correction term f(s) is tuned according to s(t) and the model of the motion primitive planning module:

[0069]

[0070] f target is the correction term function of the nonlinear part in the whole planning.

[0071] Therefore, the weight ω iThus f is set using f(s).

[0072]

[0073] The motion parameter of the healthy side collected by the rehabilitation exoskeleton module is a shoulder and elbow joint motion parameter; the shoulder and elbow joint motion parameter includes a shoulder joint adduction and abduction angle θ1, a shoulder joint flexion and extension angle θ2, a shoulder joint internal and external rotation angle θ3, an elbow joint flexion and extension angle θ4, and a spatial pose of a terminal end.

[0074] The application provides a personalized bilateral collaborative rehabilitation system based on a motion primitive, and aims to solve how to make a patient complete a rehabilitation process gradually in a training process in the prior art.

[0075] In some embodiments of the application, the rehabilitation exoskeleton module is first connected to the upper limb of the healthy side of the patient, and the patient performs rehabilitation exercise training, and the training data is transmitted to the motion primitive planning module as an input parameter;

[0076] The motion primitive planning module learns the motion habit of the healthy side of the patient, generates a motion route of the upper limb of the affected side of the patient, and the motion route has the motion characteristics of the healthy side of the patient.

[0077] The upper computer module displays a virtual training scene of the rehabilitation exercise of the upper limb of the patient with hemiplegia, real-time motion posture results of the healthy side and the affected side by using virtual reality technology, and can guide the patient to complete the rehabilitation training in combination with voice prompts.

[0078] In some embodiments of the application, as shown in Figure 2 The path planning module is used for calculating the initial position and the final position of each joint of the exoskeleton according to the initial point and the terminal point of the task issued by the upper computer module, that is, the initial pose and the expected pose of the exoskeleton are known, and then each joint variable in the two states can be obtained by inverse kinematics operation.

[0079] Taking one of the joints θ1 as an example, assuming that the initial time is t0, the relationship between the angle change of each joint in the time (t0, tf) and the time is described using a cubic polynomial, which is called a joint angle function. The initial angle is θ10, and the desired angle is θ1f. Then, the joint angle function curve of joint 1 must pass through the points (t0, θ10) and (tf, θ1f).

[0080] where the constraint conditions are

[0081] 1. The t0 position and tf position of θ1 are θ10 and θ1f;

[0082] 2. The velocity of θ1 at the t0 position and tf position is 0, i.e., the derivative of θ1 is 0;

[0083] 3. The joint acceleration of θ1 at the t0 position and tf position is 0, i.e., the derivative of the derivative of θ1 is 0;

[0084] According to the above constraint conditions, the joint angle function curve of joint 1 can be obtained.

[0085] In some embodiments of the present application, as shown in Figure 3 a DH model is established for the shoulder joint and the elbow joint of the upper limbs of the human body, and a DH model is established for each joint of the exoskeleton. The human body kinematics solving module calculates the end position coordinates of the upper limbs of the human body according to the joint coordinate positions obtained by the path planning module and the motion primitive module through forward kinematics, sends the end position coordinates of the upper limbs of the human body to the exoskeleton kinematics solving module, and then the exoskeleton kinematics solving module calculates the positions of each joint through inverse kinematics according to the end position coordinates of the upper limbs of the human body, and sends the joint positions to the controller module.

[0086] According to the disclosure and teachings of the above specification, those skilled in the art in the field of the present application can also make changes and modifications to the above embodiments. Therefore, the present application is not limited to the specific embodiments disclosed and described above, and some modifications and changes of the present application should fall within the protection scope of the claims of the present application. According to the disclosure and teachings of the above specification, those skilled in the art in the field of the present application can also make changes and modifications to the above embodiments. Therefore, the present application is not limited to the specific embodiments disclosed and described above, and some modifications and changes of the present application should fall within the protection scope of the claims of the present application.

Claims

1. A motion primitive based personalized upper limb cooperative rehabilitation system, characterized in that, The personalized upper limb cooperative rehabilitation system comprises an upper computer module, a path planning module, a motion primitive planning module, a human body kinematics solving module, an exoskeleton kinematics solving module, a controller module and a rehabilitation exoskeleton module. The rehabilitation exoskeleton module serves as an execution mechanism of the system and is used for collecting real-time motion parameters of a healthy side and assisting a hemiplegic limb to complete rehabilitation motion training. The upper computer module is used for giving a patient a task target point and interacting with the patient in rehabilitation training by using virtual reality. The path planning module is used for interpolating between a target point and a starting point and inputting to the motion primitive planning module. The human body kinematics solving module converts discrete points on a task space into discrete points on a human body joint space, and then converts the discrete points on the human body joint space into discrete points on an exoskeleton joint space through the exoskeleton kinematics solving module. The controller module drives the exoskeleton module through an actuator according to the discrete points on the exoskeleton joint space to assist the patient in motion. The model of the motion primitive planning module is as follows: wherein, is a time scaling factor, x is the position of the exoskeleton end-effector in the task space, x0and g represent the initial and target positions, respectively, v represents the velocity, represents the acceleration, K and D are the stiffness and damping parameters of the system, and f is a function of the correction term to be tuned.

2. A motion primitive based personalized upper limb cooperative rehabilitation system according to claim 1, characterized in that, The rehabilitation exoskeleton module is used for assisting a hemiplegic upper limb of a patient in rehabilitation motion training.

3. The motion primitive based personalized upper limb cooperative rehabilitation system according to claim 1, wherein, The healthy side upper limb first performs motion task training through the rehabilitation exoskeleton module, acquires real-time motion parameters of the healthy side through the rehabilitation exoskeleton module, and transmits training data to the motion primitive planning module as input parameters. The correction term function f of the motion primitive planning module uses f(s) for setting, and f(s) is as follows: where a i is a width value of a Gaussian kernel, c i is a center value of a Gaussian kernel, w i is an adjustable weight, s is a dependent variable with time t as an independent variable, and i is a serial number.

4. The motion primitive based personalized upper limb cooperative rehabilitation system according to claim 1, wherein, d(s) is a Gaussian basis function. Let the patient's healthy side first with the exoskeleton module to perform a series of rehabilitation training tasks recorded as position vector , and each time sampling point (t=0, 1, 2…T), T refers to a serial number of discrete time points; (ii) subsequently calculating a velocity vector of the trajectory in the task space from the position information and an acceleration vector ; The calculation process of the motion primitive planning module is as follows: where a is the duration of the movement, is a time scaling factor; ③Calculate s(t) is a correction term function; Therefore, the weight w can be solved by using the least square method i Thus, f is set using f(s): 。 5. The motion primitive based personalized upper limb cooperative rehabilitation system according to claim 1, wherein, ④Finally, according to s(t) and the model of the motion primitive planning module, the correction term f(s) is set.

6. A motion primitive based personalized upper limb cooperative rehabilitation system according to claim 5, wherein, The shoulder and elbow joint movement parameters include a shoulder joint adduction and abduction angle , a shoulder joint flexion and extension angle , a shoulder joint internal and external rotation angle , an elbow joint flexion and extension angle , and a spatial pose of the end.

7. A personalized upper limb coordinated rehabilitation system based on motor primitives according to claim 1, characterized in that, The motion parameters collected by the exoskeleton module are shoulder and elbow joint motion parameters.

8. A motion primitive based personalized upper limb cooperative rehabilitation system according to any one of claims 1-7, characterized in that, The path planning module is used for calculating the starting position and the final position of each joint of the exoskeleton according to the initial point and the end point of the task issued by the upper computer module through inverse kinematics, and then obtaining each joint variable in the two states through inverse kinematics operation. The working process of the human body kinematics solving module comprises: A DH model is established according to a human shoulder joint and an elbow joint.

9. A motion primitive based personalized upper limb cooperative rehabilitation system according to claim 8, characterized in that, The end position coordinates of the human body are calculated through forward kinematics according to the joint coordinates obtained by the path planning module and the motion primitive module. The working process of the exoskeleton kinematics solving module comprises: The position of each joint of the exoskeleton is calculated through inverse kinematics according to the end position coordinates obtained by the human body kinematics solving module.

Images