A rope-driven upper limb exoskeleton assistance method based on dynamic pretension and intention prediction control

The rope-driven upper limb exoskeleton assistance method using dynamic pretensioning and intention prediction control solves the problems of wire rope slack and tension fluctuation in rope-driven upper limb exoskeleton systems, achieving high-response assistance and safety controllability, and improving system stability and user experience.

CN123103593APending Publication Date: 2026-11-17JIMEI UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611337024.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-31
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing rope-driven upper limb exoskeleton systems suffer from delayed response and unstable assist output due to slack or tension fluctuations in the steel wire ropes. They also lack multi-constraint safety controls, making it difficult to achieve active response to human movement intentions and stable adjustment of rope tension.

Method used

By employing dynamic pretensioning and intention prediction control methods, and by acquiring constant angular velocity of the motor and IMU data, combined with quaternion kinematics models and extended Kalman filters, a high-precision estimation and prediction of human upper limb posture is achieved, and the tension of the steel wire rope is dynamically adjusted to construct a rope-driven upper limb exoskeleton assistance system.

Benefits of technology

It achieves high-response assist capability and safety controllability of the rope-driven upper limb exoskeleton, improves the transmission stability of the system and user wearing comfort, reduces the probability of machine failure, and realizes short-term prediction and real-time assist control of human movement intentions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123103593A_ABST
    Figure CN123103593A_ABST
Patent Text Reader

Abstract

This invention relates to the field of wearable robots and human motion assistance technology, and particularly to a rope-driven upper limb exoskeleton assistance method based on dynamic pretensioning and intention prediction control. The method includes providing a rope-driven upper limb exoskeleton, acquiring the angular velocity of a motor during constant rotation, obtaining the length of the steel wire traction rope unwound from the spiral cable storage trough per unit time, acquiring the circumference of the main tightening gear of the steel wire rope tensioner, setting tension coefficients for both unwinding and rewinding scenarios, ensuring that the total circumference of the main tightening gear of the steel wire rope tensioner during unwinding and rewinding is the same as the unwinding length of the spiral cable storage trough, and then outputting the target angular velocity of the tensioning wheel by an encoder stepper motor. This provides an upper limb exoskeleton system that combines flexible transmission characteristics, high-response assistance capability, and safe controllability, and achieves active assistance control and stable rope tension adjustment based on motion intention prediction.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of wearable robots and human motion assistance technology, and in particular to a rope-driven upper limb exoskeleton assistance method based on dynamic pretensioning and intention prediction control. Background Technology

[0002] With the increasing demand for human-machine collaborative work in fields such as intelligent manufacturing, logistics and medical rehabilitation, wearable exoskeleton robots, as an assistive device that can enhance human motor ability and reduce muscle burden, are gradually becoming an important research direction in the intersection of robotics and biomechanics. Especially in upper limb assistive scenarios, exoskeleton systems are widely used in repetitive handling, assembly operations, and rehabilitation training to improve problems such as excessive load on the human upper limbs and exercise fatigue.

[0003] Existing upper limb exoskeleton systems mainly fall into two categories: rigid linkage structures and flexible cable-driven structures. Rigid exoskeletons typically couple human joint movements directly through mechanical linkages, offering high positioning accuracy and output stability. However, due to their high structural rigidity, they often suffer from poor wearing comfort, stringent joint alignment requirements, and significant constraints on natural human movement, making them unsuitable for complex human-machine collaborative scenarios. In contrast, cable-driven exoskeleton systems use motors to drive steel cables or flexible cables to remotely traction human joints. They offer advantages such as lightweight construction, good flexibility, and high wearability, effectively reducing the overall system weight and improving user comfort.

[0004] However, existing rope-driven systems still have the following technical problems. Specifically, the wire rope is prone to slack or tension fluctuations during the winding and unwinding process, resulting in delayed system response and unstable assist output; secondly, existing systems mostly adopt a single feedback control method, relying on posture or displacement feedback for passive adjustment, making it difficult to predict and actively respond to human movement intentions in advance; frictional loss, trajectory deviation, and entanglement are prone to occur during rope transmission, affecting the long-term operational reliability of the system; and in complex human-machine interaction processes, there is a lack of effective multi-constraint safety control mechanisms, resulting in insufficient ability to suppress overload, malfunctions, and abnormal postures.

[0005] Therefore, how to construct an upper limb exoskeleton system that combines flexible transmission characteristics, high responsiveness and assistance capabilities, and safety and controllability, and achieve active assistance control based on motion intention prediction and stable adjustment of rope tension, has become a technical problem that urgently needs to be solved in this field.

[0006] It should be noted that the information disclosed in this background section is intended only to enhance the understanding of the overall background of the present invention, and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0007] This invention provides a rope-driven upper limb exoskeleton assistance method based on dynamic pretension and intention prediction control, the rope-driven upper limb exoskeleton assistance method comprising the following steps: S1: Provides a rope-driven upper limb exoskeleton, comprising a wire rope tensioner, a wire traction rope, a tension sensor, a detection device, gloves, a waist belt, a pressure sensor, a power module, and a shoulder strap. The shoulder strap and waist belt are connected by a fabric strap, and the power module and shoulder strap are attached by Velcro. The power module includes a motor and a spiral cable storage groove. The power module leads out the wire traction rope, which passes through the wire rope tensioner. The tension sensor is connected to the middle of the wire traction rope, and the end of the wire traction rope is connected to the detection device. The detection device is connected to the gloves, and the pressure sensor is mounted on the gloves. S2: Obtain the angular velocity of the motor when it is rotating at a constant speed, and obtain the linear velocity of the rope on the surface of the spiral wire storage groove and the number of rotations of the spiral wire storage groove per unit time based on the radius of the spiral wire storage groove. S3: The length of the wire traction rope unwound from the spiral wire storage trough per unit time is obtained based on the number of rotations of the spiral wire storage trough per unit time. S4: Obtain the circumference of the main tightening gear of the wire rope tensioner, and set the tension coefficient for both unwinding and winding conditions. S5: Based on the fact that the total circumference of the main tightening gear of the wire rope tensioner is the same as the unwinding length of the spiral storage groove during unwinding and rewinding, the number of rotations of the main tightening gear of the wire rope tensioner per unit time is obtained, thereby obtaining the angular velocity that the encoder stepper motor of the wire rope tensioner should be set, and then the encoder stepper motor outputs the target angular velocity of the tensioning wheel.

[0008] Furthermore, the power module also includes a cover plate, a winding wheel cover, a limiting bearing, a hollow steel wire rope guide tube, a pinion fixing cover, a cloth tape fixing component, a Teflon tube connector cover, an electrical control box cover, an electrical control box, a wire clamp cover, a wire guide, a reducer, a power bracket, wire screws, a pinion, a housing, a Teflon tube connector, a Teflon tube connector base, a guide rod, a guide screw, and a guide screw encoder. The cover plate and the winding wheel cover are fixed with screws, the spiral wire storage groove and the wire clamp cover are fixed with screws, the limiting bearing and the spiral wire storage groove are circumferentially positioned, the reducer transmits torque to the motor, the motor and the housing are connected with screws, the cloth tape fixing component and the housing are fixed with screws, the electrical control box cover is mounted on the cover plate, and the electrical control box is connected to the electrical control... The cover, the power bracket and the pinion fixing cover are fixed by screws. The pinion meshes with the outer ring gear at the bottom of the spiral wire storage groove. The pinion is circumferentially fixed with the D-shaped section at the bottom of the guide screw. The wire screw rotates spirally with the guide screw. The guide screw's circumferential rotation is restricted by the guide rod. The hollow steel wire rope guide is fixed by the wire screw. The hollow steel wire rope guide can slide and reciprocate with the hole of the pinion fixing cover. The Teflon tube connector cover and the Teflon tube connector base are connected by screws. The Teflon tube connector is connected to the Teflon tube connector cover. The wire guide is tangent to the side surfaces of the two spiral wire storage grooves. The guide screw encoder is installed at the bottom of the guide screw. The wire rope tensioner includes a tensioning bracket, a primary tensioning gear, a secondary tensioning gear, a preload spring, a wire traction rope inlet end, a bracket, and an encoder stepper motor. The primary tensioning gear rotates synchronously with the encoder stepper motor. The bracket is screwed to the encoder stepper motor. The lower end of the wire traction rope inlet end is threaded into the bracket. The tensioning bracket and the preload spring are fixed by screws. The secondary tensioning gear meshes with the primary tensioning gear. The wire traction rope includes a spiral wire storage groove lead-out end, a hollow wire rope guide end, a power module wire traction rope outlet end, a wire traction rope tensioning inlet end, and a wire traction rope tensioning outlet end. The wire traction rope is wrapped in the spiral groove of the spiral wire storage groove. The spiral wire storage groove lead-out end is led out from the tangent of the surface of the spiral wire storage groove. The wire traction rope passes through the hollow wire rope guide end. The power module wire traction rope outlet end and the wire traction rope tensioning inlet end are connected by a Teflon tube. The wire traction rope tensioning inlet end passes through the wire rope tensioner to obtain the wire traction rope tensioning outlet end.

[0009] Furthermore, the rope-driven upper limb exoskeleton assistance method also includes the following steps: S7: Acquire triaxial acceleration, angular velocity and magnetometer data from the left and right wrist IMUs and the torso IMU; S8: Based on the data obtained in step S7, the angular velocity is updated by integral using a quaternion kinematic model to obtain the attitude quaternion representation. At the same time, the accelerometer and magnetometer observations are introduced to constrain and correct the attitude. Then, under the extended Kalman filter framework, the IMU system is modeled as a nonlinear state-space model, where the state variables include the attitude quaternion and the angular velocity bias. The optimal attitude estimation is then achieved through a two-stage recursive process of prediction and update. In the prediction stage, the attitude state is predicted by time integral based on the gyroscope angular velocity. In the update stage, the attitude estimation results are observed and corrected using the gravity direction information measured by the accelerometer and the geomagnetic direction information provided by the magnetometer. S9: Using the posture of the torso IMU as the reference coordinate system for the human body, local coordinate systems are established for the left and right wrist IMUs respectively, and their quaternion or Euler angle representation in the world coordinate system is obtained through posture calculation; then, the postures of the left and right wrists are uniformly transformed to the relative coordinate system with the torso IMU as the reference through coordinate transformation relationship, thereby eliminating the influence of changes in the overall posture of the human body on the calculation of local arm movements; then, the relative posture data is normalized, and the angle range and direction are uniformly calibrated so that the postures of the left and right upper limbs are comparable and consistent under a unified reference frame. S10: Based on the relative posture information after coordinate transformation, calculate the key joint motion parameters of the left and right upper limbs of the human body. A simplified rigid body chain model of the human upper limb is constructed by using the relative posture difference between the trunk IMU and the left and right wrist IMUs. The quaternion or rotation matrix is ​​converted into Euler angles by the posture decomposition method to extract the joint angle information of each degree of freedom. At the same time, the range of joint angles is physically constrained by combining the motion constraints of the human upper limbs. S11: The posture information output by the torso IMU and the left and right wrist IMUs is fused. The data from the torso IMU and the left and right wrist IMUs are synchronized in time and aligned in space to eliminate errors caused by different sampling frequencies and communication delays. Then, a multi-sensor fusion algorithm based on extended Kalman filtering or complementary filtering is used to estimate the state of the posture information. At the same time, a redundant observation mechanism is introduced to perform weighted fusion processing on the data consistency between the left and right wrist IMUs and the torso IMU, thereby reducing the impact of single sensor drift, random noise and long-term integration error on system stability. Finally, a high-precision human upper limb posture state quantity after filtering and optimization is output.

[0010] Furthermore, the specific calculation process for steps S7-S11 is as follows: Acquire triaxial acceleration, angular velocity, and magnetometer data from the IMUs of both wrists, and define the spatial rotation state of the forearm relative to the world coordinate system using quaternions:

[0011] in, The real part of the quaternion. Let be the imaginary part of the quaternion, t be the sampling time, and T be the transpose of the matrix. ; Then convert it to a rotation matrix Expanding on:

[0012] The torso IMU acquired triaxial acceleration, angular velocity, and magnetometer data, and the spatial rotation state of the torso relative to the world coordinate system was defined using quaternions as follows:

[0013] in, The real part of the quaternion. For the imaginary part of the quaternion, ; Convert to rotation matrix Expanded to: ; Obtain the forearm orientation in the world coordinate system:

[0014] in, ; ; Using the torso coordinate system as the human body reference, all world coordinates are transformed to torso coordinates as follows:

[0015]

[0016] in, This represents the left-right component of the forearm. This represents the vertical component of the forearm. For the anteroposterior component of the forearm; Calculate the angle between the forearm and upper arm using the formula for the angle between two spatial vectors. Let a be the first spatial vector and b be the second spatial vector. Then the angle between the upper arm and the forearm is:

[0017] Depend on =1, =1, thus obtaining ( ) From the upper arm direction vector ,have to

[0018] The initial wrist angle is θ(t), and the wrist angle in the flexed state is θ(t+Δt). The flexion amplitude over time Δt is then calculated as Δθ(t) = θ(t+Δt). θ(t); Differentiating the bending amplitude Δθ(t) over time Δt yields the actual angular velocity of the bending amplitude change over time Δt. .

[0019] Furthermore, the specific calculation process for steps S2-S5 is as follows: When the human body performs an arm curl within time Δt, the length of the rope is recorded as the distance from the shoulder to the wrist. Then, based on the length of the upper arm... Forearm length The length of the rope extending out at the following different times can be calculated:

[0020]

[0021] but Inside The length of the change is:

[0022] The length of the wire rope stored in one loop of the spiral wire storage trough is:

[0023] Where P is the pitch and R is the effective radius; Based on the formula for the change in rope length over time Δt and the length of wire rope stored in one revolution of the winding sheave, the number of revolutions of the winding sheave over time Δt is:

[0024] The angular velocity of the spiral wire storage trough is the same as the angular velocity of the motor. :

[0025] The motor's output shaft is connected to a reducer, and the reducer's reduction ratio is i. Then the angular velocity of the motor's output shaft is:

[0026] The surface linear velocity of the spiral wire storage trough is obtained from its angular velocity. This surface linear velocity is also the velocity of the wire traction rope at the end of the spiral wire storage trough from which the wire traction rope is drawn out, as follows:

[0027] The velocity of the steel wire traction rope exiting the rope end is obtained based on the rope friction formula:

[0028] Where η1 is the friction coefficient, The wrap angle of the steel wire traction rope as it passes through the hollow steel wire rope conduit; When the wire rope passes between the exit end and the tensioning entry end, due to linear velocity loss, its friction coefficient is η2. Therefore, the actual output speed of the wire rope is:

[0029] After passing through the tensioner, the steel wire traction rope experiences linear velocity loss, resulting in a friction coefficient of η3. Therefore, the actual output speed of the steel wire traction rope is:

[0030] With steel wire traction ropes tied to both wrists, the actual speed exerted on the hands is... The direction of this velocity is a linear velocity perpendicular to the forearm direction. Based on the length L2 of the human forearm, the theoretical angular velocity at the wrist is obtained:

[0031] The difference between the angular velocity at the wrist measured by the IMU and the angular velocity of the wrist pulled by the motor is calculated:

[0032] The angular velocity set by the motor in different modes The angular velocity measured by the actual encoder of the motor and The angular velocities at the left and right wrists measured by the IMU, the angular velocity of the wrist pulled by the motor, and these three values ​​are used to calculate the motor's angular velocity output function through angular velocity compensation.

[0033] Wherein, K1 and K2 are both dimensionless feedback gain coefficients. K1 is the motor angular velocity tracking error compensation coefficient, which is used to perform closed-loop correction of the motor angular velocity command based on the deviation between the target angular velocity of the motor and the actual angular velocity of the motor measured by the encoder. K2 is the human-machine motion angular velocity error compensation coefficient, which is used to perform collaborative correction of the motor angular velocity command based on the deviation between the theoretical angular velocity of the wrist calculated by the rope motion model and the actual angular velocity detected by the wrist inertial measurement unit. The original output of the pressure sensor is denoted as Since pressure sensors are typically not perfectly linear, their calibration and fitting are used to obtain the calibrated palm force. :

[0034] Where a2 is the secondary calibration coefficient of the pressure sensor, used to describe the sensor's original output S P The nonlinear relationship between (t) and the actual palm force; a0 is the zero-point offset compensation coefficient; When the object is held vertically in the palm of the hand, its mass is:

[0035] When grasped from both sides of the object, its mass is:

[0036] in, ; The forearm direction vector in the torso coordinate system is The direction of unit gravity in the torso coordinate system is denoted as Then the attitude load coefficient is defined as:

[0037] Where β(t) is the angle between the forearm direction and the gravity direction, when the forearm is vertical. =0; The load torque generated by the object on the elbow joint is:

[0038] The rope is perpendicular to the forearm, and the effective lever arm of the rope on the wrist is... The assist ratio is 0≤ If ≤1, then the target assist tension required to move the object is:

[0039] The dynamic inertial force generated when an object moves with an arm is:

[0040] in, ; When the body is not carrying an object and the arm is stationary, a small preload needs to be maintained. Considering the pre-tension of the base, the weight of the object, and the dynamic movement of the arm, the final target tension is:

[0041] Measured by a tension sensor The tension error is defined as:

[0042] Among them, when The actual tension of the rope is insufficient; when At that time, the rope tension was too high; when At that time, the rope tension meets the current load requirements; The tension speed compensation amount is generated from the tension error, and PI control is used as follows:

[0043] Establish a dynamic tension coefficient, and assume the target linear velocity of the tensioning wheel is... :

[0044] in, The base tension factor is given by m, where m represents the winding or unwinding condition. This is the direction correction factor; Because the tension wheels act in different directions during winding and unwinding, they are set separately. , The dynamic tension coefficient is defined as:

[0045] The final target speed of the tensioner pulley is:

[0046] The radius of the tensioner is r, and the target angular velocity of the tensioner is: .

[0047] This invention provides a rope-driven upper limb exoskeleton assistance method based on dynamic pretensioning and intention prediction control. The power module converts the rotational motion of the motor into the winding or unwinding motion of the wire rope, thereby transmitting the wire rope motion as a linear traction force to the upper limbs. A dynamic pretensioning adjustment mechanism then adjusts the wire rope tension in real time to eliminate slack, compensate for gaps, and improve transmission stability. The method can acquire the length of the wire rope unwound from the spiral storage groove, allowing for the setting of the tension coefficient of the subsequent wire rope tensioner. This high degree of automation, requiring no user intervention, reduces the probability of machine failure. Furthermore, the fused upper limb posture information is output as a human motion state vector. Based on this vector, short-term prediction of the upper limb movement trend is achieved by analyzing the posture change rate, angular velocity, and acceleration characteristics, providing feedforward control data for the exoskeleton assistance control system. The final output posture state is simultaneously used to drive the rope-driven exoskeleton actuator, achieving real-time human-machine collaborative assistance control.

[0048] This invention, combining steps S2-S11, constitutes a rope-driven upper limb exoskeleton based on dynamic pretension and intention prediction control. First, it identifies movement conditions, including when both the upper and lower arms are extended and close to the torso, and when the upper arm is extended and close to the torso while the lower arm is bent to its limit. These conditions are based on the instantaneous flexion-lifting force exertion process of the lower arm or the long-term static maintenance of the movement. The goal is to detect the length of the steel wire traction rope during winding and unwinding. Then, based on the maximum and minimum unwinding lengths of the spiral cable storage groove, the number of motor rotations is determined to plan the safe working space of the upper limb exoskeleton. Therefore, the forearm can move freely within this range. The speed of the forearm movement is detected by the inertial measurement element of the detection device, thereby controlling the speed of the motor movement and altering the assist speed of the exoskeleton.

[0049] Other features and advantages of the present invention will be set forth in the following description, and some of the technical features and advantages may be apparent from the description or learned by practicing the invention. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, some of the drawings in the following description are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 This is a schematic flowchart of a rope-driven upper limb exoskeleton assistance method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a rope-driven upper limb exoskeleton provided in an embodiment of the present invention; Figure 3 This is an exploded structural diagram of a power module provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the power module provided in one embodiment of the present invention from another perspective; Figure 5 This is a schematic diagram of the structure of a wire rope tensioner provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of the power module rope output situation provided in an embodiment of the present invention.

[0052] Figure label: 1. Wire rope tensioner; 2. Wire rope traction rope; 3. Tension sensor; 4. Detection device; 5. Gloves; 6. Waist belt; 7. Pressure sensor; 8. Power module; 101. Tensioning bracket; 102. Main tensioning gear; 103. Secondary tensioning gear; 104. Pre-tensioning spring; 105. Wire rope traction rope inlet end; 106. Bracket; 107. Encoder stepper motor; 201. Spiral wire storage groove wire rope outlet end; 202. Hollow wire rope guide tube end; 203. Power module wire rope traction rope outlet end; 204. Wire rope traction rope tensioning inlet end; 205. Wire rope traction rope tensioning outlet end; 501. Cover Plate; 502 Winding wheel cover; 503 Limit bearing; 504 Motor; 505 Hollow steel wire rope conduit; 506 Pinion fixing cover; 507 Fabric tape fastener; 508 Teflon tube connector cover; 509 Electrical control box cover; 510 Electrical control box; 511 Wire clamp cover; 512 Spiral wire storage trough; 513 Wire guide; 514 Reducer; 515 Bracket; 516 Wire screw; 517 Pinion; 518 Housing; 519 Teflon tube connector; 520 Teflon tube connector base; 521 Guide rod; 522 Guide screw; 523 Guide screw encoder. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. The technical features designed in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0054] In the description of this invention, it should be understood that the terms "center," "lateral," "upper," "lower," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or component referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more. Additionally, the term "comprising" and any variations thereof mean "at least comprising."

[0055] Please see Figures 1 to 6 To achieve at least one or more of the aforementioned advantages, an embodiment of the present invention provides a rope-driven upper limb exoskeleton assistance method based on dynamic pretensioning and intention prediction control, the rope-driven upper limb exoskeleton assistance method comprising the following steps: S1: A rope-driven upper limb exoskeleton is provided. The rope-driven upper limb exoskeleton includes a wire rope tensioner 1, a wire traction rope 2, a tension sensor 3, a detection device 4, a glove 5, a waist belt 6, a pressure sensor 7, a power module 8, and a shoulder strap. The shoulder strap and waist belt 6 are connected by a fabric strap, and the power module 8 is attached to the shoulder strap by Velcro. The power module 8 includes a motor 504 and a spiral cable storage groove 512. The wire traction rope 2 extends from the power module 8 and passes through the wire rope tensioner 1. The tension sensor 3 is connected to the middle of the wire traction rope 2, and the end of the wire traction rope 2 is connected to the detection device 4. The detection device 4 is connected to the glove 5, and the pressure sensor 7 is disposed on the glove 5. In some embodiments, the detection device 4 can be a smart posture detection wristband, which can be integrally stitched with the glove 5. The wire traction rope 2 is broken in the middle to connect the tension sensor 3.

[0056] S2: Obtain the angular velocity of the motor 504 when it is rotating at a constant speed, and obtain the linear velocity of the rope on the surface of the spiral wire storage groove 512 and the number of rotations of the spiral wire storage groove 512 per unit time based on the radius of the spiral wire storage groove 512.

[0057] S3: The length of the wire traction rope 2 unwound by the spiral wire storage trough 512 per unit time is obtained based on the number of rotations of the spiral wire storage trough 512 per unit time.

[0058] S4: Obtain the circumference of the main tensioning gear 102 of the wire rope tensioner 1, and set the tension coefficient for the two cases of unwinding and winding.

[0059] S5: Based on the fact that the total circumference of the main tightening gear 102 of the wire rope tensioner 1 rotates per unit time during unwinding and rewinding is the same as the unwinding length of the spiral wire storage groove 512, the number of rotations of the main tightening gear 102 of the wire rope tensioner 1 per unit time is obtained, thereby obtaining the angular velocity that the encoder stepper motor 107 of the wire rope tensioner 1 should be set, and then the encoder stepper motor 107 outputs the target angular velocity of the tensioning wheel.

[0060] Furthermore, the power module 8 also includes a cover plate 501, a winding wheel cover 502, a limit bearing 503, a hollow steel wire rope conduit 505, a pinion fixing cover 506, a cloth belt fixing component 507, a Teflon tube connector cover 508, an electrical control box cover 509, an electrical control box 510, a cable clamp cover 511, a wire guide 513, a reducer 514, a power bracket 515, a wire screw 516, a pinion 517, a housing 518, a Teflon tube connector 519, a Teflon tube connector base 520, and a guide light. The system includes a rod 521, a guide screw 522, and a guide screw encoder 523. The cover plate 501 is fixed to the winding wheel cover 502 by screws. The spiral wire storage groove 512 is fixed to the wire clamp cover 511 by screws. The limit bearing 503 provides circumferential positioning to the spiral wire storage groove 512. The reducer 514 transmits torque to the motor 504. The motor 504 is connected to the housing 518 by screws. The fabric belt fastener 507 is fixed to the housing 518 by screws. The electrical control box cover 509 is located on the cover plate. On 501, the electrical control box 510 is connected to the electrical control box cover 509. The power bracket 515 is fixed to the pinion fixing cover 506 by screws. The pinion 517 meshes with the outer ring gear at the bottom of the spiral wire storage groove 512. The pinion 517 is circumferentially fixed to the D-shaped section at the bottom of the guide screw 522. The wire screw 516 rotates spirally with the guide screw 522. The guide screw 522's circumferential rotation is restricted by the guide rod 521. The hollow steel wire rope conduit 505... The wire screw 516 is fixed to the hollow steel wire rope conduit 505. The hollow steel wire rope conduit 505 can slide and reciprocate with the hole of the pinion fixing cover 506. The Teflon tube connector cover 508 and the Teflon tube connector base 520 are connected by screws. The Teflon tube connector 519 is connected to the Teflon tube connector cover 508. The wire guide 513 is tangent to the side surfaces of the two spiral wire storage grooves 512. The guide screw encoder 523 is installed at the bottom of the guide screw 522.

[0061] The wire rope tensioner 1 includes a tensioning bracket 101, a primary tensioning gear 102, a secondary tensioning gear 103, a pre-tensioning spring 104, a wire traction rope inlet end 105, a bracket 106, and an encoder stepper motor 107. The primary tensioning gear 102 rotates synchronously with the encoder stepper motor 107. The bracket 106 is screwed to the encoder stepper motor 107. The lower end of the wire traction rope inlet end 105 is threaded into the bracket 106. The tensioning bracket 101 and the pre-tensioning spring 104 are fixed by screws. The secondary tensioning gear 103 meshes with the primary tensioning gear 102.

[0062] The wire traction rope 2 includes a spiral wire storage groove lead-out wire traction rope end 201, a hollow wire rope conduit end 202, a power module wire traction rope exit end 203, a wire traction rope tensioning inlet end 204, and a wire traction rope tensioning outlet end 205. The wire traction rope 2 is wrapped in the spiral groove of the spiral wire storage groove 512. The spiral wire storage groove lead-out wire traction rope end 201 is led out from the tangent of the surface of the spiral wire storage groove 512. The wire traction rope 2 passes through the hollow wire rope conduit end 202 and the hollow wire rope conduit end 505. The power module wire traction rope exit end 203 and the wire traction rope tensioning inlet end 204 are connected by a Teflon tube. The wire traction rope tensioning inlet end 204 passes through the wire rope tensioner 1 to reach the wire traction rope tensioning outlet end 205.

[0063] Furthermore, the rope-driven upper limb exoskeleton assistance method also includes the following steps: S7: Acquire triaxial acceleration, angular velocity and magnetometer data from the left and right wrist IMUs and the torso IMU.

[0064] S8: Based on the data obtained in step S7, a quaternion kinematic model is used to integrate and update the angular velocity, resulting in an attitude quaternion representation, thus avoiding the gimbal lock problem under Euler angle representation. Simultaneously, accelerometer and magnetometer observations are introduced to constrain and correct the attitude, suppressing integral drift. Then, within the extended Kalman filter framework, the IMU system is modeled as a nonlinear state-space model, where state variables include attitude quaternions and angular velocity bias. Optimal attitude estimation is achieved through a two-stage recursive process of prediction and update. In the prediction stage, the attitude state is predicted by time integration based on the gyroscope angular velocity. In the update stage, the attitude estimation results are observed and corrected using gravity direction information measured by the accelerometer and geomagnetic direction information provided by the magnetometer, thereby improving the system's stability and noise resistance during dynamic motion.

[0065] S9: Using the posture of the torso IMU as the reference coordinate system for the human body, local coordinate systems are established for the left and right wrist IMUs respectively, and their quaternion or Euler angle representation in the world coordinate system is obtained through posture calculation; then, the postures of the left and right wrists are uniformly transformed to the relative coordinate system with the torso IMU as the reference through coordinate transformation relationship, thereby eliminating the influence of changes in the overall posture of the human body (such as leaning forward or rotating) on ​​the calculation of local arm movements; then, the relative posture data is normalized, and the angle range and direction are uniformly calibrated so that the postures of the left and right upper limbs are comparable and consistent under a unified reference frame.

[0066] S10: Based on the relative posture information after coordinate transformation, calculate the key joint motion parameters of the left and right upper limbs of the human body. A simplified rigid body chain model of the human upper limb is constructed by using the relative posture difference between the trunk IMU and the left and right wrist IMUs. The quaternions or rotation matrices are converted into Euler angles by the posture decomposition method to extract the joint angle information of each degree of freedom. At the same time, combined with the motion constraints of the human upper limbs, the range of joint angles is physically constrained to improve the biomechanical consistency of the calculation results.

[0067] S11: To improve the stability and anti-interference capability of attitude estimation, the attitude information output by the torso IMU and the left and right wrist IMUs is fused. The data from the torso IMU and the left and right wrist IMUs are time-synchronized and spatially aligned to eliminate errors caused by different sampling frequencies and communication delays. Then, a multi-sensor fusion algorithm based on extended Kalman filtering or complementary filtering is used to estimate the state of the attitude information. At the same time, a redundant observation mechanism is introduced to perform weighted fusion processing on the data consistency between the left and right wrist IMUs and the torso IMU, thereby reducing the impact of single sensor drift, random noise and long-term integration error on system stability. Finally, a high-precision human upper limb attitude state quantity after filtering and optimization is output.

[0068] Regarding the coordinate system, it can be a world inertial reference coordinate system: , where X W Y W Z W Let be the world inertial reference coordinate system, and let represent the three mutually perpendicular coordinate axes of the world coordinate system. Define the torso inertial reference coordinate system: Therefore, the vertical downward direction of the human body is: X B Y B Z B Let's define a torso coordinate system, with the three axes representing the three axes that move with the torso. Define a wrist inertial reference coordinate system: The forearm direction is: X F Y F Z FThe coordinate system is either a wrist or forearm coordinate system, representing the three coordinate axes that move with the wrist and forearm, respectively. To achieve a unified description of human upper limb movement posture and data conversion between different inertial measurement units, this invention first establishes a world inertial reference coordinate system ΣW, using this system as a unified benchmark for describing the spatial posture of the human body and exoskeleton device. Based on this world inertial reference coordinate system, a torso coordinate system and a wrist coordinate system are established according to the actual installation position and orientation of the inertial measurement units on the human torso and left and right wrists. Specifically, a torso coordinate system ΣB is established based on the torso inertial measurement module mounted on the power module 8 or the shoulder strap. This torso coordinate system moves synchronously with the human torso and serves as a reference coordinate system for analyzing human upper limb movement. A wrist coordinate system ΣF is established based on the wrist inertial measurement element mounted within the detection device 4. This wrist coordinate system moves synchronously with the human wrist and forearm and is used to describe the spatial posture changes of the forearm relative to the world coordinate system and the torso coordinate system. By uniformly calibrating the installation orientation of each inertial measurement unit, the positive directions of each coordinate axis are aligned with the anatomical direction of the human body or the longitudinal direction of the forearm, thus establishing the coordinate transformation relationship between the world coordinate system, the trunk coordinate system, and the wrist coordinate system. Based on the triaxial angular velocity, triaxial acceleration, and triaxial magnetic field data output from the trunk inertial measurement module and the wrist inertial measurement element, the control system calculates the attitude of the trunk and forearm relative to the world inertial reference coordinate system. Furthermore, it transforms the forearm direction vector from the wrist coordinate system to the trunk coordinate system to eliminate the influence of overall forward tilting, backward tilting, or rotation on the detection results of local upper limb movements. This provides a unified attitude reference for subsequent calculations of the elbow joint angle, prediction of upper limb movement intentions, and control of the steel wire traction rope 2.

[0069] Furthermore, the specific calculation process for steps S7-S11 is as follows: Acquire triaxial acceleration, angular velocity, and magnetometer data from the IMUs of both wrists, and define the spatial rotation state of the forearm relative to the world coordinate system using quaternions:

[0070] in, The real part of the quaternion. Let be the imaginary part of the quaternion, t be the sampling time, and T be the transpose of the matrix. ; Then convert it to a rotation matrix Expanding on:

[0071] The torso IMU acquired triaxial acceleration, angular velocity, and magnetometer data, and the spatial rotation state of the torso relative to the world coordinate system was defined using quaternions as follows:

[0072] in, The real part of the quaternion. For the imaginary part of the quaternion, ; Convert to rotation matrix Expanded to: ; Obtain the forearm orientation in the world coordinate system:

[0073] in, ; ; Using the torso coordinate system as the human body reference, all world coordinates are transformed to torso coordinates as follows:

[0074]

[0075] in, This represents the left-right component of the forearm. This represents the vertical component of the forearm. For the anteroposterior component of the forearm; Calculate the angle between the forearm and upper arm using the formula for the angle between two spatial vectors. Let a be the first spatial vector and b be the second spatial vector. Then the angle between the upper arm and the forearm is:

[0076] Depend on =1, =1, thus obtaining ( ) From the upper arm direction vector ,have to

[0077] The initial wrist angle is θ(t), and the wrist angle in the flexed state is θ(t+Δt). The flexion amplitude over time Δt is then calculated as Δθ(t) = θ(t+Δt). θ(t); Differentiating the bending amplitude Δθ(t) over time Δt yields the actual angular velocity of the bending amplitude change over time Δt. .

[0078] Furthermore, the specific calculation process for steps S2-S5 is as follows: When the human body performs an arm curl within time Δt, the length of the rope is recorded as the distance from the shoulder to the wrist. Then, based on the length of the upper arm... Forearm length The length of the rope extending out at the following different times can be calculated:

[0079]

[0080] but Inside The length of the change is:

[0081] The length of the wire rope stored in one turn of the spiral wire storage trough 512 is:

[0082] Where P is the pitch and R is the effective radius; Based on the formula for the change in rope length over time Δt and the length of wire rope stored in one revolution of the winding sheave, the number of revolutions of the winding sheave over time Δt is:

[0083] The angular velocity of the spiral wire storage groove 512 is the same as the angular velocity of the motor 504. :

[0084] The output shaft of motor 504 is connected to reducer 514, and the reduction ratio of reducer 514 is i. Then the output shaft angular velocity of motor 504 is:

[0085] The surface linear velocity of the spiral wire storage trough 512 is obtained based on its angular velocity. This surface linear velocity is also the velocity of the wire traction rope 2 at the end 201 of the spiral wire storage trough where the wire traction rope is led out, as follows:

[0086] The velocity of the exit end of the steel wire traction rope 2 is obtained according to the rope friction formula:

[0087] Wherein, η1 is the coefficient of friction (the steel wire traction rope 2 will have a certain frictional resistance during the movement, including the frictional force passing through the 90-degree hollow steel wire rope guide tube 505, and the coefficient of friction is η1). The angle of wrapping of the steel wire traction rope 2 after passing through the hollow steel wire rope conduit 505; During actual operation, when the wire traction rope 2 passes between the outlet end and the tensioning inlet end 204, due to linear velocity loss, its friction coefficient is η2. Therefore, the actual output speed of the wire traction rope 2 is:

[0088] During actual operation, after passing the tensioner, the steel wire traction rope 2 experiences linear velocity loss, resulting in a friction coefficient of η3. Therefore, the actual output speed of the steel wire traction rope 2 is:

[0089] With steel wire traction ropes tied to both wrists, the actual speed exerted on the hands is... The direction of this velocity is a linear velocity perpendicular to the forearm direction. Based on the length L2 of the human forearm, the theoretical angular velocity at the wrist is obtained:

[0090] The difference between the wrist angular velocity measured by the IMU and the angular velocity of the wrist pulled by the motor 504 is calculated:

[0091] The angular velocity set by motor 504 according to different modes The angular velocity measured by the actual encoder of the motor 504 and The angular velocities at the left and right wrists measured by the IMU, the angular velocity of the wrist pulled by motor 504, and the angular velocity compensation calculation of motor 504 are used to obtain the angular velocity output function of motor 504:

[0092] Wherein, K1 and K2 are both dimensionless feedback gain coefficients. K1 is the motor 504 angular velocity tracking error compensation coefficient, used to perform closed-loop correction of the motor 504 angular velocity command based on the deviation between the target angular velocity of the motor 504 and the actual angular velocity of the motor 504 measured by the encoder. K2 is the human-machine motion angular velocity error compensation coefficient, used to collaboratively correct the motor 504 angular velocity command based on the deviation between the theoretical wrist angular velocity calculated by the cable motion model and the actual angular velocity detected by the wrist inertial measurement unit. By setting K1 and K2, the weights of the motor 504's own speed tracking error and the human motion synchronization error in the motor 504 control command are adjusted respectively, thereby taking into account the response speed of the motor 504 output, the human-machine motion synchronization, and the stability of the assist process.

[0093] The original output of pressure sensor 7 is denoted as Since pressure sensor 7 is typically not perfectly linear, its calibration and fitting are used to obtain the calibrated palm force. :

[0094] Where a2 is the secondary calibration coefficient of pressure sensor 7, used to describe the sensor's original output S P (t) is a nonlinear relationship between the actual palm force and the actual palm force; a0 is the zero-point offset compensation coefficient, which represents the equivalent force offset generated by the initial preload, installation stress, temperature drift and sensor zero-point error when the pressure sensor 7 is not subjected to external force.

[0095] When the object is held vertically in the palm of the hand, its mass is:

[0096] When grasped from both sides of the object, its mass is:

[0097] in, ; The forearm direction vector in the torso coordinate system is The direction of unit gravity in the torso coordinate system is denoted as Then the attitude load coefficient is defined as:

[0098] Where β(t) is the angle between the forearm direction and the gravity direction, when the forearm is vertical. =0, the torque of the object about the elbow joint is small; when =1, the torque of the object on the elbow joint is large; The load torque generated by the object on the elbow joint is:

[0099] The rope is perpendicular to the forearm, and the effective lever arm of the rope on the wrist is... The assist ratio is 0≤ If ≤1, then the target assist tension required to move the object is:

[0100] The dynamic inertial force generated when an object moves with an arm is:

[0101] in, ; When the body is not carrying an object and the arm is stationary, a small preload needs to be maintained. Considering the pre-tension of the base, the weight of the object, and the dynamic movement of the arm, the final target tension is:

[0102] Measured by tension sensor 3 The tension error is defined as:

[0103] Among them, when The actual tension of the rope is insufficient; when At that time, the rope tension was too high; when At that time, the rope tension meets the current load requirements; The tension speed compensation amount is generated from the tension error, and PI control is used as follows:

[0104] Establish a dynamic tension coefficient, and assume the target linear velocity of the tensioning wheel is... :

[0105] in, The base tension factor is given by m, where m represents the winding or unwinding condition. This is the direction correction factor; Because the tension wheels act in different directions during winding and unwinding, they are set separately. , The dynamic tension coefficient is defined as:

[0106] The final target speed of the tensioner pulley is:

[0107] The radius of the tensioner is r, and the target angular velocity of the tensioner is: .

[0108] This invention provides a rope-driven upper limb exoskeleton assistance method based on dynamic pretensioning and intention prediction control. The power module 8 converts the rotational motion of the motor 504 into the winding or unwinding motion of the wire rope, thereby transmitting the wire rope motion as a linear traction force to the human upper limb. The dynamic pretensioning adjustment mechanism then adjusts the wire rope tension in real time to eliminate slack, compensate for gaps, and improve transmission stability. The length of the wire traction rope 2 unwound from the spiral storage groove 512 can be obtained, allowing for the setting of the tension coefficient of the subsequent wire rope tensioner 1. This highly automated process, requiring no user intervention, reduces the probability of machine failure. Furthermore, the fused upper limb posture information is output as a human motion state vector. Based on this vector, short-term prediction of the upper limb movement trend is achieved by analyzing the posture change rate, angular velocity, and acceleration characteristics, providing a feedforward control basis for the exoskeleton assistance control system. The final output posture state quantity is simultaneously used to drive the rope-driven exoskeleton actuator, achieving real-time human-machine collaborative assistance control.

[0109] This invention, combining steps S2-S11, constitutes a rope-driven upper limb exoskeleton based on dynamic pretension and intention prediction control. First, it identifies movement conditions, including when both the upper and lower arms are extended and close to the torso, and when the upper arm is extended and close to the torso while the lower arm is bent to its limit. These conditions are based on the instantaneous flexion-lifting force exertion process of the lower arm or the long-term static maintenance of the movement. The goal is to detect the length of the winding and unwinding of the steel wire traction rope 2. Then, based on the maximum and minimum unwinding lengths of the spiral cable storage groove 512, the number of rotations of the motor 504 is determined to plan the safe working space of the upper limb exoskeleton. Therefore, the forearm can move freely within this range. The speed of the forearm's movement is detected by the inertial measurement element of the detection device 4, thereby controlling the speed of the motor 504 and changing the assist speed of the exoskeleton.

[0110] Furthermore, those skilled in the art should understand that although many problems exist in the prior art, each embodiment or technical solution of the present invention can be improved in only one or a few aspects, without necessarily solving all the technical problems listed in the prior art or the background art simultaneously. Those skilled in the art should understand that any content not mentioned in a claim should not be construed as a limitation on that claim.

[0111] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A rope-driven upper limb exoskeleton assistance method based on dynamic pretensioning and intention prediction control, characterized in that: The rope-driven upper limb exoskeleton assistance method includes the following steps: S1: Provides a rope-driven upper limb exoskeleton, comprising a wire rope tensioner, a wire traction rope, a tension sensor, a detection device, gloves, a waist belt, a pressure sensor, a power module, and a shoulder strap. The shoulder strap and waist belt are connected by a fabric strap, and the power module and shoulder strap are attached by Velcro. The power module includes a motor and a spiral cable storage groove. The power module leads out the wire traction rope, which passes through the wire rope tensioner. The tension sensor is connected to the middle of the wire traction rope, and the end of the wire traction rope is connected to the detection device. The detection device is connected to the gloves, and the pressure sensor is mounted on the gloves. S2: Obtain the angular velocity of the motor when it is rotating at a constant speed, and obtain the linear velocity of the rope on the surface of the spiral wire storage groove and the number of rotations of the spiral wire storage groove per unit time based on the radius of the spiral wire storage groove. S3: The length of the wire traction rope unwound from the spiral wire storage trough per unit time is obtained based on the number of rotations of the spiral wire storage trough per unit time. S4: Obtain the circumference of the main tightening gear of the wire rope tensioner, and set the tension coefficient for both unwinding and winding conditions. S5: Based on the fact that the total circumference of the main tightening gear of the wire rope tensioner is the same as the unwinding length of the spiral storage groove during unwinding and rewinding, the number of rotations of the main tightening gear of the wire rope tensioner per unit time is obtained, thereby obtaining the angular velocity that the encoder stepper motor of the wire rope tensioner should be set, and then the encoder stepper motor outputs the target angular velocity of the tensioning wheel.

2. The rope-driven upper limb exoskeleton assistance method based on dynamic pretensioning and intention prediction control according to claim 1, characterized in that: The power module also includes a cover plate, a winding wheel cover, a limit bearing, a hollow steel wire rope guide tube, a pinion fixing cover, a fabric tape fixing component, a Teflon tube connector cover, an electrical control box cover, an electrical control box, a wire clamp cover, a wire guide, a reducer, a power bracket, wire screws, a pinion, a housing, a Teflon tube connector, a Teflon tube connector base, a guide rod, a guide screw, and a guide screw encoder. The cover plate and the winding wheel cover are fixed with screws, the spiral wire storage groove and the wire clamp cover are fixed with screws, the limit bearing and the spiral wire storage groove are used for circumferential positioning, the reducer transmits torque to the motor, the motor and the housing are connected with screws, the fabric tape fixing component and the housing are fixed with screws, the electrical control box cover is mounted on the cover plate, and the electrical control box is connected to the electrical control box cover. The power bracket and the pinion fixing cover are fixed by screws. The pinion meshes with the outer ring gear at the bottom of the spiral wire storage groove. The pinion is circumferentially fixed with the D-shaped section at the bottom of the guide screw. The wire screw rotates spirally with the guide screw. The guide screw's circumferential rotation is restricted by the guide rod. The hollow steel wire rope conduit is fixed by the wire screw. The hollow steel wire rope conduit can slide and reciprocate with the hole of the pinion fixing cover. The Teflon tube connector cover and the Teflon tube connector base are connected by screws. The Teflon tube connector is connected to the Teflon tube connector cover. The wire guide is tangent to the side surfaces of the two spiral wire storage grooves. The guide screw encoder is installed at the bottom of the guide screw. The wire rope tensioner includes a tensioning bracket, a primary tensioning gear, a secondary tensioning gear, a preload spring, a wire traction rope inlet end, a bracket, and an encoder stepper motor. The primary tensioning gear rotates synchronously with the encoder stepper motor. The bracket is screwed to the encoder stepper motor. The lower end of the wire traction rope inlet end is threaded into the bracket. The tensioning bracket and the preload spring are fixed by screws. The secondary tensioning gear meshes with the primary tensioning gear. The wire traction rope includes a spiral wire storage groove lead-out end, a hollow wire rope guide end, a power module wire traction rope outlet end, a wire traction rope tensioning inlet end, and a wire traction rope tensioning outlet end. The wire traction rope is wrapped in the spiral groove of the spiral wire storage groove. The spiral wire storage groove lead-out end is led out from the tangent of the surface of the spiral wire storage groove. The wire traction rope passes through the hollow wire rope guide end. The power module wire traction rope outlet end and the wire traction rope tensioning inlet end are connected by a Teflon tube. The wire traction rope tensioning inlet end passes through the wire rope tensioner to obtain the wire traction rope tensioning outlet end.

3. The rope-driven upper limb exoskeleton assistance method based on dynamic pretensioning and intention prediction control according to claim 2, characterized in that: The rope-driven upper limb exoskeleton assistance method also includes the following steps: S7: Acquire triaxial acceleration, angular velocity and magnetometer data from the left and right wrist IMUs and the torso IMU; S8: Based on the data obtained in step S7, the angular velocity is updated by integral using a quaternion kinematic model to obtain the attitude quaternion representation. At the same time, the accelerometer and magnetometer observations are introduced to constrain and correct the attitude. Then, under the extended Kalman filter framework, the IMU system is modeled as a nonlinear state-space model, where the state variables include the attitude quaternion and the angular velocity bias. The optimal attitude estimation is then achieved through a two-stage recursive process of prediction and update. In the prediction stage, the attitude state is predicted by time integral based on the gyroscope angular velocity. In the update stage, the attitude estimation results are observed and corrected using the gravity direction information measured by the accelerometer and the geomagnetic direction information provided by the magnetometer. S9: Using the posture of the torso IMU as the reference coordinate system for the human body, local coordinate systems are established for the left and right wrist IMUs respectively, and their quaternion or Euler angle representation in the world coordinate system is obtained through posture calculation; then, the postures of the left and right wrists are uniformly transformed to the relative coordinate system with the torso IMU as the reference through coordinate transformation relationship, thereby eliminating the influence of changes in the overall posture of the human body on the calculation of local arm movements; then, the relative posture data is normalized, and the angle range and direction are uniformly calibrated so that the postures of the left and right upper limbs are comparable and consistent under a unified reference frame. S10: Based on the relative posture information after coordinate transformation, calculate the key joint motion parameters of the left and right upper limbs of the human body. A simplified rigid body chain model of the human upper limb is constructed by using the relative posture difference between the trunk IMU and the left and right wrist IMUs. The quaternion or rotation matrix is ​​converted into Euler angles by the posture decomposition method to extract the joint angle information of each degree of freedom. At the same time, the range of joint angles is physically constrained by combining the motion constraints of the human upper limbs. S11: The posture information output by the torso IMU and the left and right wrist IMUs is fused. The data from the torso IMU and the left and right wrist IMUs are synchronized in time and aligned in space to eliminate errors caused by different sampling frequencies and communication delays. Then, a multi-sensor fusion algorithm based on extended Kalman filtering or complementary filtering is used to estimate the state of the posture information. At the same time, a redundant observation mechanism is introduced to perform weighted fusion processing on the data consistency between the left and right wrist IMUs and the torso IMU, thereby reducing the impact of single sensor drift, random noise and long-term integration error on system stability. Finally, a high-precision human upper limb posture state quantity after filtering and optimization is output.

4. The rope-driven upper limb exoskeleton assistance method based on dynamic pretensioning and intention prediction control according to claim 3, characterized in that: The specific calculation process for steps S7-S11 is as follows: Acquire triaxial acceleration, angular velocity, and magnetometer data from the IMUs of both wrists, and define the spatial rotation state of the forearm relative to the world coordinate system using quaternions: in, The real part of the quaternion. Let be the imaginary part of the quaternion, t be the sampling time, and T be the transpose of the matrix. ; Then convert it to a rotation matrix Expanding on: The torso IMU acquired triaxial acceleration, angular velocity, and magnetometer data, and the spatial rotation state of the torso relative to the world coordinate system was defined using quaternions as follows: in, The real part of the quaternion. For the imaginary part of the quaternion, ; Convert to rotation matrix Expanded to: ; Obtain the forearm orientation in the world coordinate system: in, ; ; Using the torso coordinate system as the human body reference, all world coordinates are transformed to torso coordinates as follows: in, This represents the left-right component of the forearm. This represents the vertical component of the forearm. For the anteroposterior component of the forearm; Calculate the angle between the forearm and upper arm using the formula for the angle between two spatial vectors. Let a be the first spatial vector and b be the second spatial vector. Then the angle between the upper arm and the forearm is: Depend on =1, =1, thus obtaining ( ) From the upper arm direction vector ,have to The initial wrist angle is θ(t), and the wrist angle in the flexed state is θ(t+Δt). The flexion amplitude over time Δt is then calculated as Δθ(t) = θ(t+Δt). θ(t); Differentiating the bending amplitude Δθ(t) over time Δt yields the actual angular velocity of the bending amplitude change over time Δt. .

5. The rope-driven upper limb exoskeleton assistance method based on dynamic pretensioning and intention prediction control according to claim 4, characterized in that: The specific calculation process for steps S2-S5 is as follows: When the human body performs an arm curl within time Δt, the length of the rope is recorded as the distance from the shoulder to the wrist. Then, based on the length of the upper arm... Forearm length The length of the rope extending out at the following different times can be calculated: but Inside The length of the change is: The length of the wire rope stored in one loop of the spiral wire storage trough is: in, P is the pitch, and R is the effective radius; Based on the formula for the change in rope length over time Δt and the length of wire rope stored in one revolution of the winding sheave, the number of revolutions of the winding sheave over time Δt is: The angular velocity of the spiral wire storage trough is the same as the angular velocity of the motor. : The motor's output shaft is connected to a reducer, and the reducer's reduction ratio is i. Then the angular velocity of the motor's output shaft is: The surface linear velocity of the spiral wire storage trough is obtained from its angular velocity. This surface linear velocity is also the velocity of the wire traction rope at the end of the spiral wire storage trough from which the wire traction rope is drawn out, as follows: The velocity of the steel wire traction rope exiting the rope end is obtained based on the rope friction formula: in, η1 is the coefficient of friction. The wrap angle of the steel wire traction rope as it passes through the hollow steel wire rope conduit; When the wire rope passes between the exit end and the tensioning entry end, due to linear velocity loss, its friction coefficient is η2. Therefore, the actual output speed of the wire rope is: After passing through the tensioner, the steel wire traction rope experiences linear velocity loss, resulting in a friction coefficient of η3. Therefore, the actual output speed of the steel wire traction rope is: With steel wire traction ropes tied to both wrists, the actual speed exerted on the hands is... The direction of this velocity is a linear velocity perpendicular to the forearm direction. Based on the length L2 of the human forearm, the theoretical angular velocity at the wrist is obtained: The difference between the angular velocity at the wrist measured by the IMU and the angular velocity of the wrist pulled by the motor is calculated: The angular velocity set by the motor in different modes The angular velocity measured by the actual encoder of the motor and The angular velocities at the left and right wrists measured by the IMU, the angular velocity of the wrist pulled by the motor, and these three values ​​are used to calculate the motor's angular velocity output function through angular velocity compensation. in, K1 and K2 are both dimensionless feedback gain coefficients. K1 is the motor angular velocity tracking error compensation coefficient, which is used to perform closed-loop correction of the motor angular velocity command based on the deviation between the target angular velocity of the motor and the actual angular velocity of the motor measured by the encoder. K2 is the human-machine motion angular velocity error compensation coefficient, which is used to perform collaborative correction of the motor angular velocity command based on the deviation between the theoretical angular velocity of the wrist calculated by the rope motion model and the actual angular velocity detected by the wrist inertial measurement unit. The original output of the pressure sensor is denoted as Since pressure sensors are typically not perfectly linear, their calibration and fitting are used to obtain the calibrated palm force. : Where a2 is the secondary calibration coefficient of the pressure sensor, used to describe the sensor's original output S P The nonlinear relationship between (t) and the actual palm force; a0 is the zero-point offset compensation coefficient; When the object is held vertically in the palm of the hand, its mass is: When grasped from both sides of the object, its mass is: in, ; The forearm direction vector in the torso coordinate system is The direction of unit gravity in the torso coordinate system is denoted as Then the attitude load coefficient is defined as: Where β(t) is the angle between the forearm direction and the gravity direction, when the forearm is vertical. =0; The load torque generated by the object on the elbow joint is: The rope is perpendicular to the forearm, and the effective lever arm of the rope on the wrist is... The assist ratio is 0≤ If ≤1, then the target assist tension required to move the object is: The dynamic inertial force generated when an object moves with an arm is: in, ; When the body is not carrying an object and the arm is stationary, a small preload needs to be maintained. Considering the pre-tension of the base, the weight of the object, and the dynamic movement of the arm, the final target tension is: Measured by a tension sensor The tension error is defined as: Among them, when The actual tension of the rope is insufficient; when At that time, the rope tension was too high; when At that time, the rope tension meets the current load requirements; The tension speed compensation amount is generated from the tension error, and PI control is used as follows: Establish a dynamic tension coefficient, and assume the target linear velocity of the tensioning wheel is... : in, The base tension factor is given by m, where m represents the winding or unwinding condition. This is the direction correction factor; Because the tension wheels act in different directions during winding and unwinding, they are set separately. , The dynamic tension coefficient is defined as: The final target speed of the tensioner pulley is: The radius of the tensioner is r, and the target angular velocity of the tensioner is: 。