Man-machine co-fusion type lattice structure flexible shoulder joint exoskeleton robot

Abstract

The invention discloses a man-machine co-fusion type lattice structure flexible shoulder joint exoskeleton robot which comprises two mounting bases, a lattice structure driver and a robot control system, the lattice structure flexible driver is composed of six complete driving units and four partial driving units which are all made of TPU materials, and the robot control system is arranged in the driver. All the buckling airbags are connected to the same air pressure, and all the stretching airbags are also connected to the same air pressure. The stretching air bag and the buckling air bag form an antagonism structure, and stepless switching of two states can be achieved by inflating the two air bags. The robot can assist the shoulder joint to move in the buckling direction and the stretching direction, and has the advantages of being small in size, light in weight, comfortable to wear, low in cost and the like.

Description

Technical Field

[0001] The present invention belongs to the field of upper limb exoskeleton rehabilitation robots, and in particular relates to a human-machine synergistic lattice structure flexible shoulder joint exoskeleton robot. Background Art

[0002] Stroke is a common cardiovascular disease and a leading cause of limb disability. Upper limb motor impairment caused by stroke can have a serious impact on patients' daily lives. The upper limbs are responsible for approximately 60% of the body's motor functions, and their loss of motor function creates significant inconvenience for patients in their work and personal lives. Traditional rehabilitation therapy typically requires therapists to assist with upper limb joint traction exercises, often using rigid linkage structures to assist with rehabilitation training. Currently, upper limb rehabilitation training systems are primarily categorized into two types: end-traction and exoskeleton-based rehabilitation robots. End-traction upper limb rehabilitation robots can assist with upper limb rehabilitation training and offer the advantages of ease of wear and strong adaptability, but they often suffer from joint compensation issues, limiting their effectiveness. Exoskeleton-based upper limb rehabilitation robots, on the other hand, can assist with precise movement of specific joints, but most rely on rigid linkage structures, making them difficult to wear and unsafe.

[0003] Flexible exoskeleton rehabilitation robots can effectively solve the above problems. With the advantages of low stiffness and flexibility of flexible materials, they have good wearability and safety. For example:

[0004] Patent CN109363892B proposes a rope-driven parallel flexible upper limb rehabilitation robot, which can help patients with upper limb rehabilitation training;

[0005] Patent CN111759659B proposes a portable wearable upper limb rehabilitation robot, which rotates a pulley through a drive system fixed to the waist of a double-layer clothing structure of a special clothing module. The rotation of the pulley pulls the Bowden cable core, thereby driving the human upper limbs to perform rehabilitation training exercises.

[0006] However, there are still some problems with the above-mentioned flexible upper limb rehabilitation robot patent:

[0007] 1. The wearable part of the rehabilitation device is still rigid, making it impossible to achieve overall flexibility and safety;

[0008] 2. The driving efficiency of rope-driven rehabilitation robots is low, making it difficult to assist limbs in large-scale rehabilitation training. Summary of the Invention

[0009] To solve the above problems, the present invention discloses a human-machine synergistic lattice structure flexible shoulder joint exoskeleton robot, which can assist the shoulder joint in flexion and extension movements and has the advantages of small size, light weight, comfortable wearing and low cost.

[0010] To achieve the above object, the technical solution of the present invention is as follows:

[0011] A human-machine collaborative lattice structure flexible shoulder joint exoskeleton robot includes two mounting seats, a lattice structure driver, and a robot control system.

[0012] The lattice structure driver is a foldable grid structure as a whole. The outside is a herringbone shape formed by the bottom surface of the driver and the rotating surface of the driver. The internal array arranges 10 drive units, which specifically include 6 complete drive units and 4 partial drive units.

[0013] The complete drive unit of the lattice structure driver is made of TPU material. Each complete drive unit consists of a diamond structure composed of 6 TPU materials, specifically including a peripheral support structure, an expansion airbag and a flexion airbag. Among them, the peripheral support structure is a diamond frame composed of a single layer of TPU material, and the expansion airbag and flexion airbag are respectively composed of two layers of TPU material, which can be inflated in the middle. The expansion airbag and the flexion airbag form an antagonistic structure. The expansion airbag is a double-layer TPU material with a double bend. The expansion airbag is attached to the inner wall of the peripheral support structure, and the flexion airbag is set on the diagonal of the diamond structure. One end of the flexion airbag 17 is connected to the bend of the expansion airbag.

[0014] The manufacturing process for the lattice structure actuator is as follows: First, TPU material is used to weld the peripheral support structure together to form a diamond structure. The two layers of TPU are then heat-pressed around the edges using a high-frequency machine to form a closed cavity. Next, one end of the flexion airbag is welded to the middle bend of the expansion airbag, and the other end is welded to the inner wall of the peripheral support structure. Finally, the expansion airbag is welded to the center of the peripheral support structure. Finally, six complete drive units are welded together in three layers (one, two, and three units per layer), followed by four partial drive units welded on the outside to form the lattice structure actuator.

[0015] Some units of the lattice structure driver only contain a stretchable airbag and a layer of TPU material on the outside as support. The movement angle of the driver is achieved by controlling the air pressure of the airbag, but this driver needs to be used in conjunction with other drivers to achieve bidirectional movement.

[0016] Inside the lattice structure driver, all flexion airbags and extension airbags are connected under the same air pressure and controlled separately. The lattice structure driver is connected to the upper arm through mounting bracket one and mounting bracket two, wherein the rotating surface of the driver fits with the upper arm, and the bottom surface of the driver fits with the side of the body's trunk. Then, the extension airbags and flexion airbags inside the driver unit are inflated separately to assist the patient's shoulder joint in flexion and extension movements. The angle of the lattice structure driver is determined by the angle between the bottom surface of the driver and the rotating surface of the driver. Since the extension airbag and the flexion airbag constitute an antagonistic structure, inflating the two airbags separately can achieve stepless switching between the two states.

[0017] The control system for the flexible shoulder exoskeleton robot consists of a control box and built-in control algorithms. The control box contains an air pump, an electrical proportional valve, and a control circuit board. The control circuit board outputs electrical signals that control the electrical proportional valve, thereby controlling the air pressure inside the two airbag cavities.

[0018] The flexible shoulder exoskeleton robot is based on a human-machine collaborative control method, including a top-level control strategy based on kinematics and a bottom-level control method based on model predictive control. First, the output force and bending angle of the driver are measured as system feedback values; then, based on the flexible exoskeleton control strategy of the state machine, control instructions are issued to control the motion trajectory of the flexible exoskeleton; finally, the flexible exoskeleton robot control method based on model predictive control realizes the bottom-level motion control of assisting patients' shoulder joints in smooth and safe operation, thereby assisting patients' limbs in smooth and safe movement rehabilitation.

[0019] The flexible shoulder joint exoskeleton robot adopts a top-level control strategy based on kinematics. It uses an IMU (inertial measurement unit) and force sensors to determine the angular velocity and angular acceleration of the shoulder joint, as well as the contact force between the shoulder joint and the exoskeleton robot, thereby realizing the robot's state recognition and control. The details are as follows;

[0020] First, the IMU and force sensor are installed to ensure that the directions of the IMU in the arm and torso are aligned with the directions of the humerus and spine respectively; two force sensors are installed on the inner side of the arm to stably measure the contact force between the actuator and the patient's limb; calibration is required before the system is operated, including horizontal and vertical movement of the arm, to determine the initial position of the IMU and to clear the angle, and at the same time, to clear the initial force of the force sensor to eliminate the influence of the initial contact force.

[0021] The flexible shoulder exoskeleton robot uses a state machine-based transition method to switch between the auxiliary state and the transparent state. The transition between control states is determined by the logical statements of the shoulder joint and torso kinematics. The transparent state is switched to the auxiliary state when the following conditions occur: 1) The arm elevation angle θ is greater than the starting angle θ on ; 2) Arm angular velocity is positive; 3) the initial trunk inclination angle Φ at the trunk angle Φ i Tolerance threshold Φ tol 4) The torso IMU gyroscope normal angle |ω| is lower than the maximum reading during the initial calibration; 5) The values of both sensors are greater than 20% of the maximum reference force; The state machine will transition from the auxiliary state to the transparent state when one of the following three conditions occurs: 1) The arm elevation angle is lower than the offset θ off ; 2) The elevation velocity exceeds the offset velocity Threshold; 3) The values of the two sensors are less than the maximum reference force F R During the initial calibration of the controller, the user performed a series of stretching exercises consistent with common life tasks to calibrate the state machine parameters.

[0022] The flexible shoulder exoskeleton robot adopts an underlying control method based on model predictive control. First, the system is modeled and the model parameters are estimated through a data-driven method; then, research on the model predictive control algorithm is carried out to optimally control the system.

[0023] When the limb joints move at a fixed velocity, the joint dynamics satisfy the following second-order linear system with constant coefficients:

[0024]

[0025] Where i is the number of iterations, k is the time instance, is the joint angle estimated by the model, P i (k) is the input of the system, G i is the transfer function of the system, u is the model parameter; z -1 、z -2 is the delay operator, and the system error vector E is defined as i and the model information matrix F i :

[0026]

[0027] Among them, q i is the actual joint angle, N = T / T s , T s is the sampling time; the true model parameters are searched by minimizing the error between the actual angle and the model estimated angle Among them, E i =q i -F i u i ,u i is the i-th model parameter, and the Gauss-Newton method is used to update the model parameters: ρ is the learning step size, which is a constant value. Based on the above model learning method, the dynamic model of the system is obtained:

[0028]

[0029] In order to meet the dynamic model constraints of the above system, a reference trajectory tracking method is proposed. This method is based on the model predictive control algorithm and solves the optimal solution through the quadratic programming optimization algorithm. The feedback pressure P i (k) Through internal MPC calculation:

[0030]

[0031] Where Q and R are the weight factors of the error state and control input, h and H are the rolling and prediction horizons, r(k) is the reference trajectory of the system, and y i (k) is the output of the system.

[0032] Based on the above control method, it is possible to assist patients in limb movements and improve the accuracy and safety during the movement.

[0033] The effective effects of this patent:

[0034] 1. The lattice structure flexible actuator is made of TPU material, which is a purely flexible material. It can assist the shoulder joint in flexion and extension. It has the advantages of small size, light weight, comfortable wearing and low cost.

[0035] 2. The robot adopts a state machine-based system control method, which has the advantages of easy installation and low cost. It can recognize the user's movement intention and assist the limbs in movement.

[0036] 3. The drive unit adopts an antagonistic structure and consists of two airbags. The movement directions of the two airbags repel each other, which can realize rotation control in two directions and control the output stiffness of the driver during movement.

[0037] 4. The lattice structure driver is composed of an arrangement of driving units, which can significantly improve the output force of the flexible driver without affecting the volume and mass, reduce the gas required for driving, and thus improve the response speed of the flexible driver. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 2. This is a schematic diagram of the lattice structure shoulder joint exoskeleton robot of the present invention when worn in a flexed state;

[0039] Figure 2 This is a schematic diagram of the lattice structure shoulder joint exoskeleton robot of the present invention when worn in an extended state;

[0040] Figure 3 This is a schematic diagram of the buckling state of the lattice structure driver of the present invention;

[0041] Figure 4 This is a schematic diagram of the lattice structure driver of the present invention in an extended state;

[0042] Figure 5 is a schematic diagram of a buckling state of the drive unit according to the present invention;

[0043] Figure 6 is a schematic diagram of the driving unit of the present invention in an extended state;

[0044] Figure 7 is a schematic diagram of the components of the drive unit of the present invention;

[0045] Figure 8 is a schematic diagram of a processing method of a drive unit according to the present invention;

[0046] Figure 9 Schematic diagram of the human-machine co-ordinated control method of the control system of the present invention;

[0047] Figure 10 is a schematic diagram of the sensor installation and calibration scenario of the present invention;

[0048] Figure 11 is a schematic diagram of the top-level control strategy based on the state machine of the present invention;

[0049] Figure 12 It is a schematic diagram of the underlying control method based on model predictive control described in the present invention.

[0050] List of Figure Symbols:

[0051] 1-mounting seat 1, 2-mounting seat 2, 3-upper limb arm, 4-drive unit 4-1, 5-drive unit 3-1, 6-drive unit 4-2, 7-drive unit 2-1, 8-drive unit 3-2, 9-drive unit 1-1, 10-drive unit 4-3, 11-drive unit 2-2, 12-drive unit 3-3, 13-drive unit 4-4, 14-drive bottom surface, 15-drive rotation surface, 16-extension airbag, 17-flexion airbag, 18-peripheral support structure, 19-extension airbag upper layer, 20-extension airbag lower layer, 21-flexion airbag upper layer, 22-flexion airbag lower layer. DETAILED DESCRIPTION

[0052] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention.

[0053] As shown in the figure, the human-machine collaborative lattice structure flexible shoulder joint exoskeleton robot described in the present invention includes two mounting seats, a lattice structure driver, and a robot control system.

[0054] The lattice structure driver is a foldable grid structure as a whole. The outer side is a herringbone formed by the driver bottom surface 14 and the driver rotating surface 15. The internal array has 10 drive units, which specifically include 6 complete drive units and 4 partial drive units.

[0055] The complete drive unit of the lattice structure driver is made of TPU material. Each complete drive unit consists of a diamond structure composed of 6 TPU materials, specifically including a peripheral support structure 18, an expansion airbag 16 and a flexion airbag 17. Among them, the peripheral support structure 18 is a diamond frame composed of a single layer of TPU material, and the expansion airbag 16 and the flexion airbag 17 are respectively composed of two layers of TPU material, which can be inflated in the middle. The expansion airbag 16 and the flexion airbag 17 form an antagonistic structure, wherein the expansion airbag 16 is a double-layer TPU material with a double bend. The expansion airbag 16 is attached to the inner wall of the peripheral support structure 18, and the flexion airbag 17 is set on the diagonal of the diamond structure. One end of the flexion airbag 17 is connected to the bend of the expansion airbag 16.

[0056] The manufacturing process of the lattice structure driver is as follows: first, TPU material is used to make it, and the peripheral support structure is welded together to form a diamond structure. The two layers of TPU material are then hot-pressed around the sides using a high-frequency frequency machine to form a closed cavity. Next, one end of the flexion airbag is welded to the middle bend of the expansion airbag, and the other end is welded to the connection point of the inner wall of the peripheral support structure; then, the expansion airbag is welded to the middle of the peripheral support structure; finally, 6 complete drive units are welded together in sequence (divided into three layers, with 1, 2, and 3 units in each layer respectively), and then 4 partial drive units are welded on the outside to form the lattice structure driver, such as Figure 3 and 4 shown.

[0057] Some units of the lattice structure driver only contain a stretchable airbag 16 and a layer of TPU material on the outside as support. The movement angle of the driver is achieved by controlling the air pressure of the airbag, but the driver needs to be used in conjunction with other drivers to achieve bidirectional movement.

[0058] Inside the lattice structure driver, all the flexion airbags 17 and extension airbags 16 are connected under the same air pressure and are controlled individually. The lattice structure driver is connected to the upper limb arm 3 through the mounting base 1 and the mounting base 2. Figure 2As shown, the driver rotating surface 15 is in contact with the upper arm 3, and the driver bottom surface 14 is in contact with the side of the body's trunk. Then, the extension airbag 16 and the flexion airbag 17 inside the drive unit are inflated respectively, thereby assisting the patient's shoulder joint to flex and extend. The angle of the driver is determined by the angle between the driver bottom surface 14 and the driver rotating surface 15. Since the extension airbag 16 and the flexion airbag 17 constitute an antagonistic structure, inflating the two airbags separately can achieve stepless switching between the two states. The entire shoulder joint exoskeleton robot has the advantages of small size, light weight, comfortable wearing and low cost.

[0059] The control system for the flexible shoulder exoskeleton robot consists of a control box and built-in control algorithms. The control box contains an air pump, an electrical proportional valve, and a control circuit board. The control circuit board outputs electrical signals that control the electrical proportional valve, thereby controlling the air pressure inside the two airbag cavities.

[0060] The flexible shoulder joint exoskeleton robot is based on the human-machine collaborative control method, which includes two key methods, including the top-level control strategy based on kinematics and the bottom-level control method based on model predictive control. The technical route of the control method is as follows: Figure 9 As shown in the figure, the actuator's output force and bending angle are measured as system feedback. Then, based on the flexible exoskeleton control strategy of the state machine, control instructions are issued to control the flexible exoskeleton's motion trajectory. Finally, a flexible exoskeleton robot control method based on model predictive control achieves smooth and safe underlying motion control of the patient's shoulder joint, thereby assisting the patient's limbs in smooth and safe motor rehabilitation.

[0061] The flexible shoulder joint exoskeleton robot adopts a top-level control strategy based on kinematics, using IMU (inertial measurement unit) and force sensor to determine the angular velocity and angular acceleration of the shoulder joint and the contact force between the shoulder joint and the exoskeleton robot, thereby realizing the state recognition and control of the robot; Figure 10 As shown in the figure, the IMU and force sensor are first installed to ensure that the directions of the IMU in the arm and torso are aligned with the directions of the humerus and spine respectively; the two force sensors are installed on the inner side of the arm to stably measure the contact force between the actuator and the patient's limb; calibration is required before the system is operated, including horizontal and vertical movement of the arm, to determine the initial position of the IMU and to clear the angle, and at the same time, to clear the initial force of the force sensor to eliminate the influence of the initial contact force.

[0062] The flexible shoulder joint exoskeleton robot adopts a state machine-based conversion method, and the robot switches between the auxiliary state and the transparent state. The conversion between the control states is determined based on the logical statements of the shoulder joint and torso kinematics, such as Figure 11As shown, when the following situations occur, the transparent state is switched to the auxiliary state: 1) the arm elevation angle θ is greater than the starting angle θ on ; 2) Arm angular velocity is positive; 3) the initial trunk inclination angle Φ at the trunk angle Φ i Tolerance threshold Φ tol 4) The torso IMU gyroscope normal angle |ω| is lower than the maximum reading during the initial calibration; 5) The values of both sensors are greater than 20% of the maximum reference force; The state machine will transition from the auxiliary state to the transparent state when one of the following three conditions occurs: 1) The arm elevation angle is lower than the offset θ off ; 2) The elevation velocity exceeds the offset velocity Threshold; 3) The values of the two sensors are less than the maximum reference force F R During the initial calibration of the controller, the user performed a series of stretching exercises consistent with common life tasks to calibrate the state machine parameters.

[0063] The flexible shoulder exoskeleton robot adopts the underlying control method based on model predictive control, such as Figure 12 First, the system is modeled and the model parameters are estimated using a data-driven approach. Then, a model predictive control algorithm is studied to optimally control the system.

[0064] When the limb joints move at a fixed velocity, the joint dynamics satisfy the following second-order linear system with constant coefficients:

[0065]

[0066] Where i is the number of iterations, k is the time instance, is the joint angle estimated by the model, P i (k) is the input of the system, G i is the transfer function of the system, u is the model parameter; z -1 、z -2 is the delay operator, and the system error vector E is defined as i and the model information matrix F i :

[0067]

[0068] Among them, q i is the actual joint angle, N = T / T s , T s is the sampling time; the true model parameters are searched by minimizing the error between the actual angle and the model estimated angle Among them, E i =q i -F i u i,u i is the i-th model parameter, and the Gauss-Newton method is used to update the model parameters: ρ is the learning step size, which is a constant value. Based on the above model learning method, the dynamic model of the system is obtained:

[0069]

[0070] In order to meet the dynamic model constraints of the above system, a reference trajectory tracking method is proposed. This method is based on the model predictive control algorithm and solves the optimal solution through the quadratic programming optimization algorithm. The feedback pressure P i (k) Through internal MPC calculation:

[0071]

[0072] Where Q and R are the weight factors of the error state and control input, h and H are the rolling and prediction horizons, r(k) is the reference trajectory of the system, and y i (k) is the output of the system.

[0073] Based on the above control method, it is possible to assist patients in limb movements and improve the accuracy and safety during the movement.

[0074] It should be noted that the above content merely illustrates the technical idea of the present invention and cannot be used to limit the scope of protection of the present invention. For ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications all fall within the scope of protection of the claims of the present invention.

Claims

1. A human-machine collaborative lattice structure flexible shoulder joint exoskeleton robot, characterized by: It includes a mounting base 1, a mounting base 2, a lattice structure driver, and a robot control system; The lattice structure driver is a foldable grid structure. The outer side is a herringbone formed by the bottom surface of the driver and the rotating surface of the driver. The internal array is arranged with 10 drive units, including 6 complete drive units and 4 partial drive units. The complete driving unit of the lattice structure driver is made of TPU material. Each complete driving unit consists of a diamond structure composed of 6 TPU materials, specifically including a peripheral support structure, a stretching airbag and a flexing airbag. Among them, the peripheral support structure is a diamond frame composed of a single layer of TPU material, and the stretching airbag and the flexing airbag are respectively composed of two layers of TPU material, which can be inflated in the middle. The stretching airbag and the flexing airbag constitute an antagonistic structure, among which the stretching airbag is a double-layer TPU material with a double-layer bend. The stretching airbag is attached to the inner wall of the peripheral support structure, and the flexing airbag is set on the diagonal of the diamond structure. One end of the flexing airbag is connected to the bend of the stretching airbag. Some of the driver units of the lattice structure driver only contain the inner expansion airbag and a layer of TPU material on the outer side as support, and some of the driver units are set on the outside of the complete driver unit; Inside the lattice structure actuator, all flexion and extension airbags are connected to the air source. The lattice structure actuator is connected to the upper arm through mounting brackets 1 and 2, wherein the actuator's rotating surface fits against the upper arm, and the actuator's bottom surface fits against the side of the body's trunk. The extension and flexion airbags inside the actuator unit are then inflated separately, thereby assisting the patient's shoulder joint in flexion and extension movements. The angle of the lattice structure actuator is determined by the angle between the actuator's bottom surface and the actuator's rotating surface. Because the extension and flexion airbags form an antagonistic structure, inflating the two airbags separately can achieve stepless switching between the two states. The robot control system consists of a control box and a built-in control algorithm; the control box contains an air pump, an electrical proportional valve and a control circuit board; the control circuit board outputs electrical signals to control the electrical proportional valve, thereby achieving control of the air pressure inside the two airbag cavities.

2. The human-machine symbiotic lattice structure flexible shoulder joint exoskeleton robot according to claim 1 is characterized by: The processing process of the lattice structure driver is as follows: first, TPU material is used to make it, and the peripheral support structure is welded together to form a diamond structure. The four sides of the two layers of TPU material are hot-pressed by a high-frequency frequency machine to form a closed cavity; then, one end of the flexion airbag is welded to the middle bend of the expansion airbag, and the other end is welded to the connection between the inner wall of the peripheral support structure; then, the expansion airbag is welded to the middle of the peripheral support structure respectively; finally, 6 complete drive units are welded together in sequence, and then 4 partial drive units are welded on the outside to form a lattice structure driver.

3. The human-machine collaborative lattice structure flexible shoulder joint exoskeleton robot according to claim 1, characterized in that: Based on the human-machine collaborative control method, including the top-level control strategy based on kinematics and the bottom-level control method based on model predictive control, the output force and bending angle of the actuator are measured first and used as system feedback values; Then, based on the flexible exoskeleton control strategy of the state machine, control instructions are issued to control the motion trajectory of the flexible exoskeleton; finally, based on the flexible exoskeleton robot control method of model predictive control, the underlying motion control of the patient's shoulder joint is achieved to assist in smooth and safe movement rehabilitation of the patient's limbs.

4. The human-machine collaborative lattice structure flexible shoulder joint exoskeleton robot according to claim 3 is characterized by: The kinematics-based top-level control strategy utilizes an IMU and force sensor to determine the angular velocity and acceleration of the shoulder joint, as well as the contact force between the shoulder joint and the exoskeleton robot, thereby enabling state recognition and control of the robot. The IMU and force sensor are first installed, ensuring that the IMU orientations of the arm and torso are aligned with the humerus and spine, respectively. Two force sensors are installed on the inside of the arm to provide stable measurement of the contact force between the actuator and the patient's limb. Calibration is required before the system is operated, including horizontal and vertical movement of the arm, determining the initial position of the IMU and clearing the angle, and clearing the initial force of the force sensor to eliminate the influence of the initial contact force.

5. The human-machine symbiotic lattice structure flexible shoulder joint exoskeleton robot according to claim 4 is characterized by: The robot switches between the auxiliary state and the transparent state using a state machine-based conversion method; the conversion between control states is determined based on the logical statements of the shoulder joint and torso kinematics. When the following situations occur, the transparent state is switched to the auxiliary state: 1) The arm elevation angle θ is greater than the starting angle θ on ; 2) Arm angular velocity is positive; 3) the initial trunk inclination angle Φ at the trunk angle Φ i Tolerance threshold Φ tol 4) The torso IMU gyroscope normal angle |ω| is lower than the maximum reading during the initial calibration; 5) The values of both sensors are greater than 20% of the maximum reference force; The state machine will transition from the auxiliary state to the transparent state when one of the following three conditions occurs: 1) The arm elevation angle is lower than the offset θ off ; 2) The elevation velocity exceeds the offset velocity Threshold; 3) The values of the two sensors are less than the maximum reference force F R 10%; the controller is calibrated initially.

6. The human-machine symbiotic lattice structure flexible shoulder joint exoskeleton robot according to claim 3, characterized in that: The underlying control method based on model predictive control is as follows: First, the system is modeled and the model parameters are estimated using a data-driven approach. Then, research on model predictive control algorithms is conducted to optimally control the system. When the limb joints move at a fixed velocity, the joint dynamics satisfy the following second-order linear system with constant coefficients: Where i is the number of iterations, k is the time instance, is the joint angle estimated by the model, P i (k) is the input of the system, G i is the transfer function of the system, u is the model parameter; Z -1 、z -2 is the delay operator, and the system error vector E is defined as i and the model information matrix F i : Among them, q i is the actual joint angle, N = T / T s , T s is the sampling time; the true model parameters are searched by minimizing the error between the actual angle and the model estimated angle Among them, E i =q i -F i u i ,u i is the i-th model parameter, and the Gauss-Newton method is used to update the model parameters: ρ is the learning step size, which is a constant value. Based on the above model learning method, the dynamic model of the system is obtained: In order to meet the dynamic model constraints of the above system, a reference trajectory tracking method is proposed. This method is based on the model predictive control algorithm and solves the optimal solution through the quadratic programming optimization algorithm. The feedback pressure P i (k) Through internal MPC calculation: Where Q and R are the weight factors of the error state and control input, h and H are the rolling and prediction horizons, r(k) is the reference trajectory of the system, and y i (k) is the output of the system; Based on the above control method, it is possible to assist patients in limb movements and improve the accuracy and safety during the movement.

Images