Method for self-adapting adjustment of astronaut exoskeleton movement intensity and intelligent trajectory planning

Abstract

The application provides an astronaut exoskeleton motion strength self-adaptive adjustment and trajectory intelligent planning method, relates to the technical field of astronaut auxiliary motion equipment, and comprises the following steps: collecting motion data to generate a characteristic vector, identifying a characteristic phase point to set a torque control envelope, dynamically calculating an angular momentum compensation value based on angular acceleration variation, and adjusting a torque limit interval by monitoring joint mobility; meanwhile, an expected trajectory is compared with an actual trajectory, a deformable virtual channel and a nonlinear damping field are constructed, inertial force interference is effectively inhibited, and stable and accurate motion control of astronauts in a weightless environment is ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of astronaut auxiliary motion equipment, and particularly relates to an astronaut exoskeleton motion intensity self-adaptive adjustment and trajectory intelligent planning method. BACKGROUND

[0002] With the normalization of space station missions and the development of future deep space exploration activities, astronauts face severe challenges in long-term work and life in space environments. In the microgravity environment, physiological problems such as muscle atrophy and bone density decline of astronauts are increasingly prominent, which threatens the health and mission execution capability of astronauts. As a new type of man-machine integration device, exoskeleton technology can effectively assist astronauts in space to exercise and perform daily tasks, and has become a research hotspot in the fields of space medicine and aerospace technology.

[0003] The existing exoskeleton system lacks self-adaptive adjustment capability for the motion characteristics of astronauts in the microgravity environment. The control algorithm developed in the ground environment cannot effectively cope with the motion characteristics dominated by inertial force in space, resulting in safety hazards such as mismatched assistive torque and insufficient joint protection, and cannot achieve precise regulation and control of the motion intensity of astronauts. Traditional trajectory planning methods mostly use rigid constraint mode, which cannot adapt to the motion uncertainty of astronauts in the microgravity environment. This fixed trajectory mode lacks dynamic adjustment mechanism, and when the astronaut deviates from the predetermined trajectory, it will produce excessive correction torque, which not only affects the motion comfort, but also causes joint stress concentration and increases the risk of injury. The existing exoskeleton system does not fully consider the joint inertia coupling effect in the microgravity environment. In the space environment, due to the lack of gravity, the motion coupling between joints is more complex, and the existing single-joint independent control strategy is difficult to effectively suppress the disturbance caused by inertial force, resulting in increased motion instability, affecting the motion effect and safety of astronauts. SUMMARY

[0004] The astronaut exoskeleton motion intensity self-adaptive adjustment and trajectory intelligent planning method provided by the embodiments of the present application can solve the problems in the prior art.

[0005] In a first aspect, the astronaut exoskeleton motion intensity self-adaptive adjustment and trajectory intelligent planning method provided by the embodiments of the present application comprises:

[0006] The motion data of the astronaut is collected in real time by the sensor of the exoskeleton device to generate a motion state feature vector;

[0007] Identify feature phase points in the movement process based on the motion state feature vector, and set a torque control envelope according to the motion features of each feature phase point; dynamically calculate an angular momentum compensation value by analyzing the instantaneous change rate of joint angular acceleration and superimpose it on the torque control envelope, while establishing a torque limit interval, dynamically adjusting the upper and lower limits of the torque limit interval by real-time monitoring of joint range of motion, and ensuring that the output torque is within a safe range;

[0008] Generate an expected motion trajectory based on the motion state feature vector, compare the expected motion trajectory with the actual motion trajectory, and obtain a trajectory deviation value;

[0009] Select multiple dynamic control points from the trajectory deviation value and perform spline interpolation to generate a deformable virtual channel to limit the movement range; calculate the inertia coupling effect in the joint movement process, and suppress the interference of inertial force by adjusting the shape and stiffness characteristics of the deformable virtual channel, construct a nonlinear damping field based on joint coupling stiffness, and adaptively adjust the damping adjustment parameters according to the relative motion acceleration between joints to dissipate excess inertial energy, ensuring that astronauts maintain a stable motion trajectory in a weightless environment.

[0010] Identifying feature phase points in the movement process based on the motion state feature vector, and setting a torque control envelope according to the motion features of each feature phase point includes:

[0011] Obtain the time domain features and frequency domain features of the motion state feature vector, and establish a motion cycle feature map based on the time domain features and the frequency domain features;

[0012] Identify feature phase points based on the motion cycle feature map, identify acceleration sudden change points by calculating the angular acceleration change rate, identify angle extreme points by analyzing angular velocity zero-crossing points, identify electromyographic signal peak points by the sliding window method, and combine the acceleration sudden change points, the angle extreme points and the electromyographic signal peak points to obtain a feature phase point set;

[0013] For each feature phase point in the feature phase point set, calculate the angle deviation between the joint angle vector corresponding to the feature phase point and the expected angle vector, and the angular velocity deviation between the angular velocity vector corresponding to the feature phase point and the expected angular velocity vector, calculate the target torque value at the feature phase point based on the angle deviation and the angular velocity deviation; based on the motion features of each feature phase point, set the weight coefficient of the feature phase point, combine the weight coefficient with the target torque value, and use a cubic spline interpolation method to interpolate the target torque value to obtain a torque control envelope.

[0014] The instantaneous change rate of the joint angle acceleration is analyzed, the angular momentum compensation value is dynamically calculated and superimposed on the torque control envelope, and a torque limit interval is established, and the upper and lower limits of the torque limit interval are dynamically adjusted by real-time monitoring of the joint range of motion, including:

[0015] The instantaneous change rate of the joint angle acceleration is calculated, the time-domain integral of the angular acceleration based on the instantaneous change rate is calculated to obtain the angular momentum, the instantaneous change of the angular momentum is combined with the deviation of the angular momentum relative to the expected angular momentum to obtain an angular momentum compensation value, and the angular momentum compensation value is superimposed on the torque output by the torque control envelope to obtain a superimposed torque;

[0016] The ratio of the superimposed torque to the maximum bearing torque of the joint is calculated to obtain a torque utilization rate, a reference safety factor is set based on the torque utilization rate, and the superimposed torque is added and subtracted by the reference safety factor to obtain the upper and lower limits of the torque limit interval.

[0017] A plurality of dynamic control points are selected from the trajectory deviation value and spline interpolation is performed to generate a deformable virtual channel to limit the motion range, including:

[0018] The local curvature of the trajectory deviation value is calculated, and the extreme points of the trajectory deviation value, the extreme points of the local curvature, and the points where the rate of change of velocity exceeds a preset velocity threshold are selected as initial dynamic control points;

[0019] The difference between the current joint range of motion and the reference range of motion is calculated to obtain a joint range of motion deviation value, the difference between the current motion speed and the reference speed is calculated to obtain a speed deviation value, the joint range of motion deviation value and the speed deviation value are added to obtain a total adjustment amount, and the total adjustment amount is superimposed on the position of the initial dynamic control point to obtain an adjusted dynamic control point.

[0020] The adjusted dynamic control points are subjected to cubic B-spline interpolation to obtain a center trajectory of the deformable virtual channel, and the width adjustment amount of the deformable virtual channel is calculated based on the joint range of motion deviation value and the speed deviation value, and the width adjustment amount is expanded along both sides of the center trajectory to form a dynamic boundary of the deformable virtual channel.

[0021] The inertia coupling effect during joint motion is calculated, and the interference of inertial force is suppressed by adjusting the shape and stiffness characteristics of the deformable virtual channel, including:

[0022] The moment of inertia parameters and coupling inertia parameters of the joint are obtained and constructed into a spatial inertia tensor matrix, and based on the spatial inertia tensor matrix and the joint angular velocity vector, a coupling inertia force is obtained.

[0023] The local deformation of the deformable virtual channel is calculated based on the amplitude of the coupled inertial force, and the local deformation is superimposed along the normal vector direction of the deformable virtual channel to obtain the deformed virtual channel.

[0024] Collect the spatial attitude information and angular attitude information of the joint to obtain the joint attitude matrix. Use the ratio of the amplitude of the coupled inertial force to the preset maximum allowable inertial force as the stiffness adjustment factor. Multiply the stiffness adjustment factor by the quadratic form of the joint attitude matrix to obtain the stiffness increment matrix of the deformable virtual channel.

[0025] The compensation torque is calculated based on the coupled inertial force. The position and shape of the deformed virtual channel are adjusted according to the compensation torque. A constraint force is applied to the deformed virtual channel based on the stiffness increment matrix to achieve dynamic suppression of inertial force interference.

[0026] Constructing a nonlinear damping field based on joint coupling stiffness, and adaptively adjusting the damping parameters according to the relative motion acceleration between joints to dissipate excess inertial energy includes:

[0027] Obtain the joint configuration vector, calculate the generalized stiffness matrix based on the joint configuration vector, and perform eigenvalue decomposition on the generalized stiffness matrix to obtain the coupling stiffness eigenvector and coupling stiffness eigenvalue.

[0028] A nonlinear damping matrix is ​​established based on the coupled stiffness eigenvector. Damping adjustment parameters are constructed based on the coupled stiffness eigenvalue and the exponential decay term of the joint motion velocity. The damping adjustment parameters are set as the coefficients of the nonlinear damping matrix to obtain the nonlinear damping field.

[0029] The relative acceleration between adjacent joints is calculated. The linear term coefficient of the damping adjustment parameter is updated based on the deviation between the amplitude of the relative acceleration and the coupling stiffness characteristic value relative to a preset stiffness threshold. The damping adjustment parameter is updated based on the amplitude of the relative acceleration and the joint motion velocity. The updated damping adjustment parameter is substituted into the nonlinear damping field. The damping torque applied by the nonlinear damping field is used to dissipate excess inertial energy during motion in a directional manner.

[0030] A second aspect of the present invention provides an adaptive adjustment system for the motion intensity of an astronaut exoskeleton and an intelligent trajectory planning system, comprising:

[0031] The first unit is used to collect astronauts’ motion data in real time through the sensors of the exoskeleton device and generate motion state feature vectors.

[0032] The second unit is configured to identify feature phase points in the motion process based on the motion state feature vector, set a torque control envelope according to the motion features of each feature phase point, dynamically calculate an angular momentum compensation value by analyzing the instantaneous change rate of the joint angle acceleration, and superimpose the angular momentum compensation value on the torque control envelope, while establishing a torque limit interval, dynamically adjusting the upper and lower limits of the torque limit interval by real-time monitoring of the joint range of motion, and ensuring that the output torque is within a safe range.

[0033] The third unit is configured to generate an expected motion trajectory based on the motion state feature vector, compare the expected motion trajectory with an actual motion trajectory, and obtain a trajectory deviation value.

[0034] The fourth unit is configured to select a plurality of dynamic control points from the trajectory deviation value and perform spline interpolation to generate a deformable virtual channel to limit the motion range, calculate the inertia coupling effect in the joint motion process, and suppress the interference of inertial force by adjusting the shape and stiffness characteristics of the deformable virtual channel, construct a nonlinear damping field based on the joint coupling stiffness, and adaptively adjust the damping adjustment parameters according to the relative motion acceleration between joints to dissipate excess inertial energy, and ensure that the astronaut maintains a stable motion trajectory in a weightless environment.

[0035] A third aspect of the embodiment of the present application,

[0036] An electronic device is provided, comprising:

[0037] A processor;

[0038] A memory for storing processor-executable instructions;

[0039] The processor is configured to invoke the instructions stored in the memory to perform the method described above.

[0040] A fourth aspect of the embodiment of the present application,

[0041] A computer-readable storage medium is provided, which stores computer program instructions, and the computer program instructions are executed by a processor to implement the method described above.

[0042] The beneficial effects of the present application are as follows:

[0043] The present application realizes precise adaptive adjustment of the exoskeleton motion intensity by real-time acquisition and intelligent analysis of the motion data of astronauts, effectively avoids the risk of joint damage in a weightless environment, and ensures motion stability.

[0044] The application constructs a nonlinear damping field and a deformable virtual channel based on joint coupling stiffness, can effectively suppress the inertial force interference in the weightlessness environment, dynamically adjusts the moment limit and damping parameters, and realizes smooth transition and precise control in the astronaut movement process.

[0045] The adaptive moment adjustment and trajectory planning method of the application solves the problem of unstable movement of the traditional exoskeleton in the weightlessness environment, improves the comfort and safety of the astronauts in the space station for movement training through angular momentum compensation and inertial coupling effect calculation, and provides technical support for the maintenance of physical health in long-term space missions. BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 1 The figure is a flowchart of the astronaut exoskeleton movement intensity adaptive adjustment and trajectory intelligent planning method of the embodiment of the application.

[0047] Figure 2 The figure is a flowchart of the inertial coupling effect suppression and virtual channel adjustment. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical scheme and advantages of the embodiment of the application clearer, the technical scheme in the embodiment of the application will be described clearly and completely in combination with the drawings in the embodiment of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor belong to the protection scope of the application.

[0049] The technical scheme of the application will be described in detail in specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in some embodiments.

[0050] Figure 1 The figure is a flowchart of the astronaut exoskeleton movement intensity adaptive adjustment and trajectory intelligent planning method of the embodiment of the application, as shown in Figure 1 The method comprises:

[0051] The motion data of the astronaut is collected in real time by the sensor of the exoskeleton device to generate a motion state feature vector;

[0052] The feature phase points in the movement process are identified based on the motion state feature vector, the moment control envelope is set according to the motion features of each feature phase point, the angular momentum compensation value is dynamically calculated by analyzing the instantaneous change rate of the joint angular acceleration and is superimposed on the moment control envelope, the moment limit interval is established, the upper and lower limits of the moment limit interval are dynamically adjusted by real-time monitoring of the joint range of motion, and the output moment is ensured to be in a safe range.

[0053] generate an expected motion trajectory based on the motion state feature vector, compare the expected motion trajectory with an actual motion trajectory, and obtain a trajectory deviation value;

[0054] select a plurality of dynamic control points from the trajectory deviation value and perform spline interpolation to generate a deformable virtual channel to limit the motion range; calculate inertia coupling effects in the joint motion process, and suppress the interference of inertial force by adjusting the shape and stiffness characteristics of the deformable virtual channel, construct a nonlinear damping field based on joint coupling stiffness, adaptively adjust damping adjustment parameters according to the relative motion acceleration between joints to dissipate excess inertial energy, and ensure that the astronaut maintains a stable motion trajectory in a weightless environment.

[0055] In an optional embodiment, the torque control envelope is set according to the motion characteristics of each feature phase point based on the motion state feature vector, including:

[0056] obtain time domain features and frequency domain features of the motion state feature vector, and establish a motion cycle feature map based on the time domain features and the frequency domain features;

[0057] identify feature phase points based on the motion cycle feature map, identify acceleration mutation points by calculating angular acceleration change rates, identify angle extreme points by analyzing angular velocity zero-crossing points, identify electromyographic signal peak points by a sliding window method, and combine the acceleration mutation points, the angle extreme points, and the electromyographic signal peak points to obtain a feature phase point set;

[0058] For each feature phase point in the feature phase point set, calculate an angle deviation between a joint angle vector corresponding to the feature phase point and an expected angle vector, and an angular velocity deviation between an angular velocity vector corresponding to the feature phase point and an expected angular velocity vector, calculate a target torque value at the feature phase point based on the angle deviation and the angular velocity deviation; set a weight coefficient of the feature phase point based on the motion characteristics of each feature phase point, combine the weight coefficient with the target torque value, and use a cubic spline interpolation method to interpolate the target torque value to obtain a torque control envelope.

[0059] The time domain features of the motion state feature vector are obtained, including angle, angular velocity, angular acceleration, and electromyographic signal. The time domain features can be obtained by wearing a sensor, such as installing an inertial measurement unit (IMU) at the joint position of the human body to collect angle and angular velocity data, and obtaining angular acceleration data by first-order differentiation of angular velocity. At the same time, surface electromyographic electrodes are placed at the relevant muscle positions to collect electromyographic signals. For example, during the rehabilitation of the knee joint, electromyographic electrodes can be placed on the front and back of the thigh to collect electromyographic signals of the quadriceps femoris and the hamstrings. The sampling frequency is set to 200 Hz to ensure the accuracy of the data.

[0060] The time domain signal is converted to the frequency domain using fast Fourier transform to extract the main frequency component, harmonic component, and energy distribution feature. For example, for normal walking motion, the main frequency of the hip joint angle signal is usually between 0.8-1.2 Hz, while during fast running, it rises to 2.0-2.5 Hz. In actual application, the hip joint flexion and extension motion of an astronaut was analyzed, and it was found that the main frequency of the motion was 1.05 Hz, and the energy was mainly concentrated in the 0.9-1.2 Hz frequency band, indicating that this was a typical normal gait cycle.

[0061] Based on the obtained time domain and frequency domain features, a motion cycle feature map is established, which includes an angle- angular velocity phase plane, an angular velocity- angular acceleration phase plane, and an electromyographic signal- angle correlation graph. In actual application, the data of an astronaut performing 10 consecutive knee joint flexion and extension motions was recorded, and by constructing the angle- angular velocity phase plane, the repeatability pattern of the motion was clearly displayed, and each cycle formed a closed loop, and the shape and size of the loop represented the stability and intensity of the motion.

[0062] After the motion cycle feature map is established, the feature phase point recognition is performed, the acceleration mutation point is recognized by calculating the angular acceleration change rate, the angular acceleration change rate is defined as the difference between the angular acceleration at adjacent two sampling points and the sampling time interval, when the absolute value of the change rate exceeds the preset threshold (for example, 150 degrees / second 3 ), it is marked as an acceleration mutation point. In a certain test, two main acceleration mutation points were detected in a motion cycle, located at the beginning of flexion and extension, and the acceleration change rates were 175 degrees / second 3 and-190 degrees / second 3 , respectively.

[0063] The angle extreme point is identified by analyzing the angular velocity zero-crossing point. The point where the angular velocity changes from positive to negative or from negative to positive is the zero-crossing point, which corresponds to the maximum or minimum angle. To improve the identification accuracy, a dead zone of ±0.5 degrees / second is introduced. That is, only when the absolute value of angular velocity is less than 0.5 degrees / second and the direction changes, it is determined as a zero-crossing point. In the test, two angle extreme points in the knee flexion and extension process are successfully identified, corresponding to the fully flexed state (about 110 degrees) and the fully extended state (about 0 degrees) respectively.

[0064] The electromyography signal peak point is identified by the sliding window method. The window width is set to 200 ms. If there is a point in the window whose amplitude is the maximum and exceeds 150% of the average value, it is marked as an electromyography signal peak point. In the knee flexion and extension movement, the quadriceps femoris electromyography signal appears a peak about 100 ms before the extension starts (amplitude is 310% of the baseline), and the hamstrings electromyography signal appears a peak about 80 ms before the flexion starts (amplitude is 275% of the baseline).

[0065] The above acceleration mutation point, angle extreme point and electromyography signal peak point are combined to form a set of characteristic phase points. For each characteristic phase point in the set, the angle deviation of its corresponding joint angle vector from the expected angle vector and the angular velocity deviation of its corresponding angular velocity vector from the expected angular velocity vector are calculated. The expected angle and the expected angular velocity can be standard data of healthy people or target values of rehabilitation training. For example, in the knee flexion and extension movement, the actual angle of the characteristic phase point (flexion start point) is 5 degrees, the expected angle is 0 degrees, and the angle deviation is 5 degrees; the actual angular velocity is 2 degrees / second, the expected angular velocity is 0 degrees / second, and the angular velocity deviation is 2 degrees / second.

[0066] Based on the calculated angle deviation and angular velocity deviation, the target torque value at the characteristic phase point is calculated. The torque calculation considers the compensation torque of the angle deviation and the damping torque of the angular velocity deviation. For example, for a characteristic phase point with an angle deviation of 5 degrees and an angular velocity deviation of 2 degrees / second, the calculated target torque value is 0.8 Nm, and the direction is the assist extension direction.

[0067] Considering the motion characteristics of each characteristic phase point, its weight coefficient is set. The weight coefficient reflects the importance of the point in the entire movement. The angle extreme point weight is set to 1.0, the acceleration mutation point weight is set to 0.8, and the electromyography signal peak point weight is set to 0.7. The weighted target torque value is obtained by combining the weight coefficient with the target torque value.

[0068] The weighted target torque value is interpolated by using a cubic spline interpolation method to obtain a complete torque control envelope. The envelope ensures smooth transition of the torque between the characteristic phase points and avoids sudden changes in the control torque. In practical applications, the torque control envelope generated for an astronaut provides an auxiliary torque of 0.6 Nm at the initial flexion stage, gradually decreases to 0 Nm at the middle flexion stage, provides an auxiliary torque of 0.8 Nm at the initial extension stage, and gradually decreases to 0.2 Nm at the completion of the extension stage. The torque changes smoothly throughout the process, effectively assisting the astronaut to complete the training action.

[0069] In an optional embodiment, the angular momentum compensation value is dynamically calculated by analyzing the instantaneous change rate of the joint angular acceleration, and is superimposed on the torque control envelope, and a torque limit interval is established, and the upper and lower limits of the torque limit interval are dynamically adjusted by real-time monitoring of the joint range of motion, including:

[0070] The instantaneous change rate of the joint angular acceleration is calculated, the time-domain integral of the angular acceleration is calculated based on the instantaneous change rate to obtain an angular momentum, the instantaneous change amount of the angular momentum is combined with the deviation of the angular momentum from the expected angular momentum to obtain an angular momentum compensation value, and the angular momentum compensation value is superimposed on the torque output by the torque control envelope to obtain a superimposed torque;

[0071] The ratio of the superimposed torque to the maximum bearing torque of the joint is calculated to obtain a torque utilization rate, a reference safety factor is set based on the torque utilization rate, and the superimposed torque is added and subtracted by the reference safety factor to obtain the upper and lower limits of the torque limit interval.

[0072] The joint angular acceleration data of the exoskeleton is obtained, and the joint angular acceleration is directly measured by a micro acceleration sensor arranged at the joint or indirectly calculated based on the angular velocity change rate of a high-precision encoder. Taking the lower limbs exoskeleton of an astronaut as an example, the knee joint and the hip joint are both equipped with an encoder with a precision of 0.01 degrees, and the sampling frequency is 500 Hz, which can detect an angular acceleration change of 0.05 degrees / s 2 .

[0073] The calculation process of the instantaneous change rate of the joint angular acceleration involves dividing the difference between the angular accelerations in adjacent time windows by the time interval. A sliding time window of 25 ms is used, the angular acceleration values at the current time t and the previous time t-25 ms are recorded, and the instantaneous change rate is obtained by dividing the difference between the two values by 0.025 seconds. For example, during the astronaut's walking process, if the knee joint angular acceleration at time t is 4.8 degrees / s 2 , and the angular acceleration at time t-25 ms is 4.2 degrees / s 2 , then the instantaneous change rate is 24 degrees / s 3 . This value reflects the degree of sudden change in the astronaut's joint movement state and is an important indicator for judging the change in movement intensity.

[0074] The angular acceleration values in the sampling period are arranged in time sequence, and the angular momentum is calculated by trapezoidal integration method. The angular acceleration value of each sampling point is multiplied by the time interval and the average value of adjacent points, and then all interval results are accumulated. In the microgravity environment, the angular acceleration values of the astronaut's hip joint within 120 ms are 1.5, 1.8, 2.2, 2.4, and 2.6 degrees per second 2 , and the sampling interval is 30 ms. The angular momentum calculated by the trapezoidal integration method is 0.585 degrees per second.

[0075] The determination of the expected angular momentum is based on the astronaut's motion intention recognition and the preset motion mode. The astronaut's motion intention is recognized according to the bioelectric signal and force sensor data, and the ideal angular momentum value at each time is calculated in advance in combination with the preset gait library. The actual measured angular momentum is compared with the expected angular momentum to calculate the deviation value. Taking the walking task in the microgravity environment as an example, the preset angular momentum of the astronaut's hip joint is 0.55 degrees per second, the actually calculated angular momentum is 0.585 degrees per second, and the deviation is 0.035 degrees per second.

[0076] The calculation of the angular momentum compensation value comprehensively considers the instantaneous change and the deviation. The instantaneous change of the angular momentum is multiplied by the weight coefficient a (the typical value in the microgravity environment is 0.55), the deviation of the angular momentum relative to the expected value is multiplied by the weight coefficient β (the typical value in the microgravity environment is 0.45), and the two are added to obtain the comprehensive compensation value. In the above example, if the instantaneous change of the angular momentum is 0.06 degrees per second, the deviation is 0.035 degrees per second, a = 0.55, and β = 0.45, then the angular momentum compensation value is 0.049 degrees per second.

[0077] The reference output torque is obtained from the torque control envelope according to the current state of the joint, and the final superimposed torque is obtained by superimposing the angular momentum compensation value. If the torque value obtained from the torque control envelope of the astronaut's knee joint is 10.5 Nm, and the angular momentum compensation value is converted to torque as 0.85 Nm, then the superimposed torque is 11.35 Nm.

[0078] The torque utilization rate is the ratio of the superimposed torque to the maximum bearing torque of the exoskeleton joint. This index reflects the percentage of the current output torque to the exoskeleton limit capacity, and is an important parameter for evaluating the system load state and safety margin. If the maximum bearing torque of the exoskeleton knee joint is 25 Nm, and the superimposed torque is 11.35 Nm, then the torque utilization rate is 45.4%.

[0079] The setting of the reference safety factor is based on dynamic adjustment of the torque utilization rate, and considering the special needs of astronauts in a microgravity environment, a specific piecewise function relationship is set: when the torque utilization rate is less than 25%, the reference safety factor is 2.0 Nm; when the torque utilization rate is between 25% and 60%, the reference safety factor is 1.5 Nm; when the torque utilization rate is higher than 60%, the reference safety factor is 0.8 Nm. In the above example, the torque utilization rate is 45.4%, falling into the 25%-60% interval, and the corresponding reference safety factor is 1.5 Nm.

[0080] The upper and lower limits of the torque limit interval are determined by adding and subtracting the superimposed torque from the reference safety factor, and the upper limit is equal to the superimposed torque plus the reference safety factor, and the lower limit is equal to the superimposed torque minus the reference safety factor. Continuing the above example, the superimposed torque is 11.35 Nm, and the reference safety factor is 1.5 Nm, so the upper limit of the torque limit interval is 12.85 Nm, and the lower limit is 9.85 Nm. Limiting the actual output torque within this interval ensures that the exoskeleton can provide sufficient assistance and will not cause harm to the astronaut due to excessive torque.

[0081] Joint range of motion monitoring is an important basis for adjusting the torque limit interval, and the range of motion is quantified by calculating the ratio of the current angle of the exoskeleton joint to its range of motion. For the astronaut's hip joint, if the range of motion is -30 degrees to +120 degrees and the current angle is 60 degrees, the range of motion is 60%. When the range of motion exceeds 80%, the upper limit of the torque limit interval is automatically reduced by 15% to ensure that the joint will not exceed the physiological limit due to excessive torque; when the range of motion is less than 20%, the lower limit of the torque limit interval is increased by 12% to prevent the joint from shaking or being unstable near the limit position.

[0082] During real-time monitoring, the joint range of motion data is updated at a frequency of 200 Hz, and the dynamic adjustment of the torque limit interval is performed every 5 ms to ensure that the control system can respond to changes in the astronaut's motion state in a timely manner. The dynamic adjustment algorithm also considers physiological indicators such as the astronaut's heart rate and breathing rate. When the heart rate exceeds 75% of the set threshold, the upper limit of the torque limit interval will be additionally reduced by 8% to avoid excessive assistance leading to insufficient exercise intensity of the astronaut.

[0083] In the activity scene inside the space station cabin, this method significantly improves the assistance effect of the exoskeleton. Taking the task of transferring experimental equipment as an example, the astronaut needs to carry an experimental box with a mass of 15 kg in a microgravity environment. The traditional control method lags behind when the astronaut's motion intention changes, while after using this method, the torque limit interval is dynamically adjusted, and the exoskeleton assistance is self-adaptively adjusted, successfully reducing the astronaut's energy consumption by 32% and shortening the task completion time by 25%.

[0084] The angular momentum compensation mechanism performs well in the simulation of a gravity environment. In a ground test, in a simulation of a lunar gravity environment (about 1 / 6 of the gravity of the Earth), when an astronaut performs a squatting action, the knee joint torque fluctuation under a traditional control method can reach ±3.5 Nm, while after using the method, due to the fast response of the angular momentum compensation value to the load change, the torque fluctuation is reduced to ±1.2 Nm, and the astronaut's subjective comfort score is increased from 7.2 to 9.1 (full score 10).

[0085] The system parameter configuration is determined by optimizing a large amount of astronaut training data. In an offline stage, the model is trained using motion data of astronauts of different body types to determine initial values of the weight coefficients α and β. In an online operation stage, the parameter values are fine-tuned according to individual differences and adaptive changes of the astronauts. Experiments show that in a walking task in a microgravity environment, the configuration of α = 0.55 and β = 0.45 can provide the best assistance effect, while in a squatting task in a simulated gravity environment, the configuration of α = 0.6 and β = 0.4 is more appropriate.

[0086] The exoskeleton intelligent trajectory planning function and the torque self-adaptive adjustment work cooperatively. According to an environment map in a space station and task requirements, an optimal motion path is planned in advance, and the assistance force is dynamically adjusted during the execution of the task by the astronaut in combination with a real-time torque control strategy. Measurement data show that after using the method, the battery endurance time of the exoskeleton is extended by 28% when the astronaut completes daily work in the cabin, and the physiological energy consumption of the astronaut is reduced by 35%, which greatly improves the efficiency of the space task and the comfort of the astronaut.

[0087] In an alternative embodiment, a plurality of dynamic control points are selected from the trajectory deviation values and are subjected to spline interpolation to generate a deformable virtual channel to limit the motion range, which includes:

[0088] The local curvature of the trajectory deviation values is calculated, and extreme points of the trajectory deviation values, extreme points of the local curvature, and points with a speed change rate exceeding a preset speed threshold are selected as initial dynamic control points;

[0089] A joint activity deviation value is obtained by calculating the difference between a current joint activity and a reference activity, a speed deviation value is obtained by calculating the difference between a current motion speed and a reference speed, the joint activity deviation value and the speed deviation value are added to obtain a total adjustment amount, and the total adjustment amount is added to the positions of the initial dynamic control points to obtain adjusted dynamic control points;

[0090] The adjusted dynamic control points are subjected to cubic B-spline interpolation to obtain a center trajectory of the deformable virtual channel, and a width adjustment amount of the deformable virtual channel is calculated based on the joint activity deviation value and the speed deviation value, and the width adjustment amount is expanded along both sides of the center trajectory to form a dynamic boundary of the deformable virtual channel.

[0091] The sliding window method was used to calculate the local curvature of each point on the trajectory. The window size was 5 sampling points. For each data point within the window, the rate of change of the angle formed by three adjacent points was calculated to obtain the estimated local curvature value of that point. When astronauts perform squatting and standing movements, the local curvature of the knee joint trajectory increases significantly at the movement transition point, with a maximum value of 0.08 / mm, while the local curvature during the steady movement phase usually remains below 0.01 / mm.

[0092] Local maxima and minima are identified on the trajectory deviation curve, as these points typically correspond to changes in the astronaut's motion state. The criteria for determining local extrema are: the trajectory deviation value at the current point is greater than or less than the deviation values ​​of the three points before and after it. During intravehicular walking maneuvers, a typical trajectory deviation curve can identify 5-8 extrema, corresponding to key moments in the gait cycle, such as heel strike and toe lift.

[0093] Identify the peaks and troughs on the local curvature curve; these points typically indicate locations where the trajectory direction changes significantly. When an astronaut completes a rotation, the local curvature of the hip joint trajectory forms distinct peaks at the start and end points of the rotation, with the maximum curvature value reaching 0.12 / mm, which is 5-6 times the average curvature.

[0094] The rate of change of joint motion velocity is calculated in real time, and when the rate of change exceeds a preset threshold, the point is marked as a velocity mutation point. Different velocity thresholds are set for different joints and different types of movements. For example, when an exoskeleton assists astronauts in performing upper limb manipulation tasks, the velocity change rate threshold for the elbow joint is set to 4 degrees / second. 2 The threshold for the rate of change of shoulder joint velocity is set to 3 degrees / second. 2 When an astronaut rapidly extends their arm to grasp an object, the rate of change of velocity at the elbow joint can reach 6 degrees per second. 2 This mutation point will be captured as a dynamic control point.

[0095] The initial dynamic control points are formed by combining the three types of feature points mentioned above. Trajectory deviation extreme points, local curvature extreme points, and velocity change points are merged. If feature points of different types are too close (less than a preset merging threshold, usually 2% of the total trajectory length), one of them is retained to avoid the control points being too dense. When an astronaut completes a standard stretching motion, 8-12 initial dynamic control points are usually generated, with an average interval of about 10% of the total trajectory length.

[0096] The joint range of motion is calculated as the ratio of the current joint angle to its maximum range of motion, each joint has its physiological limit of motion range, and the current level of activity is calculated by real-time monitoring of joint angle. Taking the knee joint as an example, assuming its range of motion is 0-140 degrees, and the current angle is 70 degrees, the range of motion is 50%. The reference range of motion is the expected range of motion value preset according to the standard motion pattern, which reflects the activity level that the joint should reach in a specific task. When the astronaut performs the squat training, the reference range of motion curve of the knee joint changes with time, gradually increasing to 80% in the squatting stage and decreasing to 20% in the standing stage.

[0097] The joint range of motion deviation is obtained by calculating the difference between the current range of motion and the reference range of motion, which reflects the deviation degree between the actual motion state and the expected state of the astronaut. Positive value indicates that the current range of motion is higher than the reference value, and negative value indicates that it is lower than the reference value. When assisting the astronaut to perform the vertical climbing task, if the current range of motion of the hip joint is 65%, and the reference range of motion is 60%, the range of motion deviation is 5%, indicating that the astronaut's hip joint activity is slightly higher than expected.

[0098] The instantaneous speed is calculated by dividing the position difference of adjacent sampling points by the time interval, and compared with the reference speed, which is the ideal speed curve preset according to the astronaut's body characteristics and task requirements. In the extravehicular activity training, the reference speed of the wrist joint is 12 degrees / second, and if the actual monitored speed is 14 degrees / second, the speed deviation is 2 degrees / second, indicating that the astronaut's hand movement is slightly faster than expected.

[0099] The calculation of speed deviation takes into account the characteristic differences of different joints, for large joints (such as hip joint, knee joint), the relative deviation rate (percentage difference between actual speed and reference speed) is used; for small joints (such as finger joints), the absolute deviation value is directly used. When the astronaut performs fine operation task, if the actual speed of the thumb joint is 3.5 degrees / second, and the reference speed is 3.0 degrees / second, the speed deviation is 0.5 degrees / second.

[0100] The total adjustment amount is calculated by considering the range of motion deviation and the speed deviation, multiplying the range of motion deviation value by the weight coefficient γ (typical value is 0.6), multiplying the speed deviation value by the weight coefficient δ (typical value is 0.4), and adding them to get the total adjustment amount. When the astronaut performs the in-cabin assembly task, if the elbow joint range of motion deviation value is 4%, and the speed deviation value is 1.5 degrees / second, after unit conversion and weight superposition, the total adjustment amount is 6.5mm.

[0101] The total adjustment amount calculated is superimposed on the initial control point position along the trajectory normal direction to form the adjusted dynamic control point, and the adjustment direction is determined by the positive and negative of the activity deviation value: positive value adjusts outward, negative value adjusts inward. In the astronaut arm lifting action, if a certain initial control point is located at a specific position of the shoulder joint trajectory, the total adjustment amount is 8 mm, and the activity deviation is positive, then the control point will move 8 mm outward to form the adjusted control point.

[0102] The adjusted dynamic control point sequence is input to generate a smooth and continuous curve through a cubic B-spline interpolation algorithm, which is the center trajectory of the deformable virtual channel. This algorithm ensures the continuity of the first and second derivatives of the curve at the control points, avoiding sudden points on the trajectory. In actual implementation, a cubic B-spline curve is constructed for every four adjacent control points, and the complete center trajectory is formed after connecting all the curve segments. For an astronaut to complete a standard rotation action, there are about 10 dynamic control points for the waist joint trajectory, and the total length of the center trajectory generated by the cubic B-spline interpolation is about 120 cm, with a curve smoothness score of 92% (full score 100%).

[0103] The adjustment amount of the channel width is calculated based on the joint activity deviation value and the speed deviation value. In specific calculation, the absolute value of the joint activity deviation value is multiplied by the coefficient μ (value range 0.15-0.25, typical value 0.2), and the absolute value of the speed deviation value is multiplied by the coefficient v (value range 0.1-0.2, typical value 0.15). The sum of the two is the width adjustment reference value, which is multiplied by the safety factor (typical value 1.2) to get the final width adjustment amount. When an astronaut is performing fine operation training, if the wrist joint activity deviation absolute value is 3% and the speed deviation absolute value is 0.8 degrees per second, the calculated width adjustment amount is 12 mm.

[0104] The dynamic boundary of the deformable virtual channel is formed by expanding on both sides of the center trajectory. The channel boundary is constructed along the normal direction of the center trajectory with the width adjustment amount as the radius. The channel width is not constant, but dynamically changes according to the local characteristics of each point along the trajectory, appropriately widening at key points of motion and narrowing at stable stages. For an astronaut to walk, the average width of the knee joint trajectory virtual channel in the swing phase is 35 mm, and the average width in the support phase is 25 mm, with a maximum width of 45 mm at the phase transition point.

[0105] The motion limiting function of the virtual channel is realized by combining soft and hard constraints. When the trajectory of the exoskeleton joint approaches the boundary of the channel (the distance is less than 20% of the width of the channel), a flexible resistance is applied, and the resistance is inversely proportional to the distance from the boundary. When the trajectory tries to cross the boundary, the resistance is immediately increased to form a hard constraint to prevent crossing the boundary. When the astronaut is training with weight, if the trajectory of the hip joint deviates too much and is about to touch the boundary of the virtual channel, a maximum correction torque of 5 N·m will be applied to guide the joint back to the safe area.

[0106] The real-time updating of the virtual channel ensures the dynamic adaptability of the system. The joint state and trajectory deviation are recalculated at a frequency of 50 Hz, the dynamic control points are updated every 200 ms, and the virtual channel is regenerated every 500 ms. This multi-level updating mechanism ensures timely response to changes in the astronaut's motion state and avoids excessive consumption of computing resources. When the astronaut performs long-term in-cabin activities, the dynamic updating of the virtual channel improves the exoskeleton assistance effect score from 7.8 points in the traditional method to 9.3 points (full score 10 points), and the astronaut's subjective comfort is improved by 35%.

[0107] In an alternative embodiment, the inertia coupling effect during joint motion is calculated, and the interference of the inertia force is suppressed by adjusting the shape and stiffness characteristics of the deformable virtual channel, comprising:

[0108] Obtain the moment of inertia parameters and coupled inertia parameters of the joint and construct a spatial inertia tensor matrix. Based on the spatial inertia tensor matrix and the joint angular velocity vector, the coupled inertia force is obtained;

[0109] Calculate the local deformation of the deformable virtual channel according to the amplitude of the coupled inertia force, and superimpose the local deformation along the normal vector direction of the deformable virtual channel to obtain the deformed virtual channel;

[0110] Collect the spatial attitude information and angle attitude information of the joint to obtain a joint attitude matrix. The ratio of the amplitude of the coupled inertia force to the preset maximum allowed inertia force is taken as a stiffness adjustment factor. The stiffness adjustment factor is multiplied by the quadratic form of the joint attitude matrix to obtain a stiffness increment matrix of the deformable virtual channel;

[0111] Based on the coupled inertia force, a compensation torque is calculated, the position and shape of the deformed virtual channel are adjusted according to the compensation torque, and a constraint force is applied to the deformed virtual channel based on the stiffness increment matrix, to realize dynamic suppression of the inertia force interference.

[0112] As shown in Figure 2 the method comprises:

[0113] The inertial coupling effect significantly affects the motion accuracy and stability of the astronaut exoskeleton system in microgravity environment. The inertia force generated by joint motion will be transmitted to adjacent joints, forming complex coupling interference. Obtaining accurate inertia parameters is the basis for suppressing the inertia coupling interference. The three-dimensional inertia scanning technology is used to accurately measure the mass distribution and inertia characteristics of each part of the exoskeleton. For the standard astronaut lower limb exoskeleton, the typical value of the main shaft inertia of the thigh part is 0.42 kg·m 2 , the lower leg part is 0.28 kg·m 2 , and the foot is 0.08 kg·m 2 .

[0114] The coupling inertia parameters reflect the mutual influence between different joint motions. The cross-inertia data during joint motion is collected by a multi-axis torque sensor. During astronaut walking, the coupling inertia parameter between the hip joint and the knee joint is about 0.15 kg·m 2 , and the coupling inertia parameter between the knee joint and the ankle joint is about 0.08 kg·m 2 . These parameters are adjusted according to the exoskeleton configuration and astronaut body size, and are automatically calibrated before each use to ensure the accuracy of the parameters.

[0115] The construction process of the spatial inertia tensor matrix involves the organization and arrangement of the main shaft inertia and the coupling inertia. The inertia values of the three main shaft directions are placed on the diagonal line of the matrix, and the coupling inertia parameters are placed on the non-diagonal line to form a 3×3 symmetric matrix. Taking the astronaut upper limb exoskeleton as an example, in the spatial inertia tensor matrix of the shoulder joint, the main diagonal elements are 0.35, 0.38, and 0.32 kg·m 2 , representing the inertia of the three main shaft directions; the non-diagonal elements are 0.12, 0.09, and 0.07 kg·m 2 , representing the coupling inertia between different shafts.

[0116] The joint angular velocity vector is obtained by high-precision gyroscopes. A three-axis gyroscope with a sampling frequency of 1000 Hz is installed at the key joints of the exoskeleton to monitor the joint motion state in real time. When the astronaut performs a body rotation action, the angular velocity vector of the waist joint has typical values of [0.52, 0.38, 0.15] rad / s, representing the angular velocity in the three main shaft directions.

[0117] The calculation of the coupling inertia force considers the spatial inertia tensor and the angular velocity vector. The angular velocity vector is multiplied by the spatial inertia tensor matrix, and then the cross product of the angular velocity vector is obtained, which is the coupling inertia force generated by joint motion. When the astronaut performs a rapid arm lifting action, the coupling inertia force generated by the shoulder joint can reach 2.8 N·m, and the main action direction is perpendicular to the motion plane. This force will cause the motion trajectory to deviate from the expected path by about 12 mm.

[0118] The compensation deformation amount required by the virtual channel is calculated according to the amplitude of the coupled inertial force, and the calculation process adopts a nonlinear mapping relationship: when the inertial force is less than 1 N·m, the deformation amount is proportional to the inertial force, and the proportional coefficient is 5 mm / (N·m); when the inertial force is between 1-3 N·m, the growth rate of the deformation amount decreases, and a square root relationship is adopted; when the inertial force is greater than 3 N·m, the deformation amount tends to be saturated, and the maximum is not more than 25 mm. Taking the astronaut's extravehicular maintenance training as an example, when the coupled inertial force generated by the wrist joint is 1.5 N·m, the calculated local deformation amount is 9.3 mm.

[0119] The normal vector of the virtual channel points to the vertical direction of the channel center trajectory, and the normal vector of each point is calculated based on the tangent direction of the virtual channel center trajectory. In three-dimensional space, the normal vector is not unique, and the direction perpendicular to the current motion plane is selected as the main normal vector, so that the interference of the inertial force can be most effectively offset. For the walking action of the astronaut in the horizontal plane, the normal vector of the knee joint virtual channel mainly points to the vertical direction.

[0120] The superposition of the local deformation amount follows the spatial distribution principle, and the calculated local deformation amount is applied to the virtual channel along the normal vector direction. The deformation amount is Gaussian distributed in space, and the deformation center is located at the action point of the inertial force, which gradually decays to both sides. When the astronaut performs a rapid squatting action, the deformation amount of the knee joint virtual channel at the motion inflection point is the largest, reaching 15 mm, and gradually decreases to 3 mm within a range of 50 mm to both sides.

[0121] The deformed virtual channel has an adaptive geometry. Through the superposition of the local deformation amount, the originally regular virtual channel forms an asymmetric shape, and the convex part is used to compensate the interference of the inertial force. When the astronaut performs a 360-degree rotating action, the hip joint virtual channel protrudes outward by about 18 mm in the acceleration stage of rotation and protrudes inward by about 20 mm in the deceleration stage, effectively compensating for the influence of centrifugal force and centripetal force.

[0122] Joint space posture information is obtained through multi-dimensional sensor fusion. A three-axis accelerometer, a three-axis gyroscope, and a three-axis magnetometer are integrated, and sensor data is fused through a Kalman filter algorithm to estimate the three-dimensional space posture of the joint in real time. In a microgravity environment, sensor data fusion pays special attention to gyroscope information, and reduces the weight of the accelerometer to improve the accuracy of posture estimation. For astronauts during extravehicular activities in the space station, the joint space posture change can be tracked with an accuracy of 0.5 degrees.

[0123] The angular posture information describes the position of the joint in the anatomical plane. The joint angle is measured by a high-precision encoder with a resolution of 0.01 degrees. In the standard rehabilitation training of astronauts, the angle range of the knee joint is 0-120 degrees, and the abduction / adduction range of the hip joint is -30 to 45 degrees. These data form the basis of the angular posture information.

[0124] The joint posture matrix combines the spatial posture and angular posture information, combining the three-dimensional spatial posture with the anatomical plane angle to construct a 3x3 posture matrix, which fully describes the spatial position and angular state of the joint. When astronauts perform compound movements, the typical value of the shoulder joint posture matrix contains the combined information of 60 degrees of flexion, 30 degrees of abduction, and 15 degrees of external rotation.

[0125] The stiffness adjustment factor reflects the degree of influence of inertial force on the system. The ratio of the measured coupled inertia force amplitude to the preset maximum allowed inertia force is taken as the stiffness adjustment factor. The preset maximum allowed inertia force is set according to the joint type and task requirements, for example, 4 N·m for the wrist joint, 8 N·m for the elbow joint, and 15 N·m for the knee joint. When astronauts perform high-speed movements, if the measured coupled inertia force of the knee joint is 9 N·m and the preset maximum allowed value is 15 N·m, the stiffness adjustment factor is 0.6.

[0126] The stiffness increment matrix is calculated by the quadratic form of the stiffness adjustment factor and the joint posture matrix. The posture matrix is multiplied by its transpose to form a quadratic form, and then multiplied by the stiffness adjustment factor to obtain the stiffness increment matrix. This matrix describes the stiffness value that needs to be increased in different directions of the virtual channel. For astronauts performing stretching training, the stiffness increment matrix of the wrist joint shows that the stiffness is increased by 30% in the flexion / extension direction and by 15% in the ulnar / radial deviation direction.

[0127] The calculation of the compensation torque is based on the direction and size of the coupled inertia force. According to Newton's third law, a compensation torque equal in size and opposite in direction to the coupled inertia force is generated to neutralize the inertial interference. When astronauts perform rapid head rotation, the coupled inertia force generated by the neck joint is 1.2 N·m, and the calculated compensation torque is -1.2 N·m, acting in the opposite direction of the inertia force.

[0128] The position and shape adjustment of the virtual channel after deformation is a dynamic process. According to the trend of the compensation torque, the inertial interference at the next moment is predicted, and the position and shape of the virtual channel are adjusted in advance. The adjustment frequency is 200 Hz, ensuring that it can track the rapidly changing motion state. In the agile training of astronauts with frequent direction changes, the shape of the elbow joint virtual channel is updated every 50 ms, and the maximum position adjustment amplitude reaches ±15 mm.

[0129] The constraint of the stiffness increment matrix on the virtual channel is reflected in the elastic property. The elastic parameters of the virtual channel are modified according to the stiffness increment matrix, so that the channel becomes more rigid in the direction of the inertial force and remains flexible in other directions. When the astronauts perform fine operation tasks, the radial stiffness of the finger joint virtual channel is increased to 2.5 times the original value, and the longitudinal stiffness is only increased by 20%, forming a directional constraint characteristic.

[0130] The application of the constraint force adopts a virtual spring-damper model, and the virtual channel is regarded as an elastic body with variable stiffness. When the joint motion deviates from the center of the channel, a compound constraint force proportional to the deviation distance and proportional to the deviation velocity is generated. When the astronauts perform weight training, if the hip joint deviates from the center of the virtual channel by 8mm, the moving speed is 50mm / s, and the generated constraint force is 3.2N, of which the elastic component is 2.4N and the damping component is 0.8N.

[0131] The dynamic suppression effect of the inertial force interference is evaluated by the motion smoothness, and the acceleration standard deviation of the joint trajectory is calculated as the smoothness index. The smaller the value is, the smoother the motion is. When the astronauts perform routine in-cabin activities, after the inertial compensation function is enabled, the acceleration standard deviation of the knee joint motion trajectory is reduced from 0.85m / s 2 to 0.32m / s 2 , with an improvement of 62%.

[0132] The actual test data shows that this method significantly improves the performance of the exoskeleton system in the microgravity environment. In the simulation of space station extravehicular operation tasks, after the inertial coupling compensation is enabled, the operation accuracy of the astronauts is improved by 48%, the task completion time is shortened by 35%, and the energy consumption is reduced by 27%. The subjective evaluation of the astronauts shows that the stability and comfort scores are improved from 7.2 to 9.5 (full score 10), especially in rapid motion and direction change, the effect of suppressing inertial interference is most obvious.

[0133] In an alternative embodiment, a nonlinear damping field based on joint coupling stiffness is constructed, and the excess inertial energy is dissipated by adaptively adjusting the damping adjustment parameter according to the relative motion acceleration between the joints, including:

[0134] An joint configuration vector is obtained, and a generalized stiffness matrix is calculated based on the joint configuration vector. The coupling stiffness eigenvector and the coupling stiffness eigenvalue are obtained by eigenvalue decomposition of the generalized stiffness matrix;

[0135] A nonlinear damping matrix is established based on the coupling stiffness eigenvector, and a damping adjustment parameter is constructed according to an exponential decay term of the coupling stiffness eigenvalue and the joint motion velocity. The damping adjustment parameter is set as the coefficient of the nonlinear damping matrix to obtain the nonlinear damping field;

[0136] The relative acceleration between adjacent joints is calculated, the linear term coefficient of the damping adjustment parameter is updated according to the deviation of the amplitude of the relative acceleration from the preset stiffness threshold value relative to the coupling stiffness eigenvalue, the damping adjustment parameter is updated according to the amplitude of the relative acceleration and the joint motion speed, the updated damping adjustment parameter is substituted into the nonlinear damping field, and the directional dissipation of the redundant inertia energy in the motion process is realized by using the damping torque applied by the nonlinear damping field.

[0137] The joint configuration vector is basic data describing the current posture of the exoskeleton, and contains the angle information of all joints. The joint angle data is collected by high-precision encoders and organized into a configuration vector. Taking a six-degree-of-freedom upper limb exoskeleton as an example, the joint configuration vector contains three angles of the shoulder joint (pitch, internal and external rotation, and abduction and adduction), the elbow joint angle, and two angles of the wrist joint (flexion and extension, and rotation). When the astronaut performs an extravehicular maintenance task, the typical configuration vector is [35°, 20°, 15°, 80°, 30°, 5°], which represents the current angles of the joints respectively.

[0138] The generalized stiffness matrix reflects the relationship between the change of joint position and the change of corresponding torque. Based on the joint configuration vector, the generalized stiffness matrix is constructed by calculating the torque change caused by a small change in joint angle. For the upper limb exoskeleton, the generalized stiffness matrix is a 6x6 symmetric matrix, and the diagonal elements represent the stiffness of each joint itself, and the non-diagonal elements represent the coupling stiffness between joints. In actual measurement, the typical value of the stiffness of the shoulder joint in the pitch direction is 2.5 Nm / rad, and the coupling stiffness between the shoulder joint and the elbow joint is about 0.8 Nm / rad. The generalized stiffness matrix is not constant and will adjust with the change of exoskeleton configuration and load.

[0139] The generalized stiffness matrix is subjected to eigenvalue decomposition to obtain a set of eigenvectors and corresponding eigenvalues. The eigenvectors represent the stiffness principal directions, and the eigenvalues represent the stiffness sizes in these directions. For the astronaut's upper limb exoskeleton, eigenvalue decomposition usually obtains 6 eigenvectors and eigenvalues. The typical eigenvalue distribution is [3.2, 2.8, 2.1, 1.7, 1.2, 0.8] Nm / rad, and the corresponding eigenvectors represent six mutually orthogonal stiffness principal directions. The direction of the largest eigenvalue is usually consistent with the main force direction of the astronaut.

[0140] The coupling stiffness eigenvector constitutes the principal axis direction of the nonlinear damping field, and a damping coordinate system is established based on these eigenvectors, so that the damping action is applied along the stiffness principal direction. This design makes the damping force consistent with the inherent stiffness characteristics of the system, and can effectively suppress vibration in a specific direction. For the task of the astronaut operating a precision instrument in a microgravity environment, the eigenvector in the direction of the largest stiffness approximately points to the normal of the operation plane, and a larger damping will be applied in this direction to ensure the operation stability.

[0141] The nonlinear damping matrix is the core of the damping field construction. The eigenvectors are used as the basis of the damping matrix, and the nonlinear damping coefficients are introduced on this basis. The damping matrix is also a 6x6 matrix, and its structure is similar to the generalized stiffness matrix, but the element value reflects the damping coefficient in each direction. Unlike traditional constant damping, the coefficients in the nonlinear damping matrix will change dynamically with the motion state, realizing intelligent damping control.

[0142] The damping adjustment parameter controls the strength of the nonlinear damping field. The damping adjustment parameter is constructed based on the coupling stiffness eigenvalue and the joint motion speed. In specific calculation, the eigenvalue is first divided by the preset stiffness reference value (usually 2.0 Nm / rad) to obtain the normalized stiffness value, and then the exponential decay term of joint speed is introduced: when the joint speed is below the threshold value (typical value is 10 degrees / s), the exponential term is close to 1, and the damping remains at the basic level; when the speed exceeds the threshold value, the exponential term decreases rapidly, and the damping decreases accordingly, avoiding excessive damping to limit normal motion. When the astronaut performs a rapid arm extension motion, the elbow joint speed reaches 40 degrees / s, and the speed exponential decay term decreases to 0.25, causing the damping strength to automatically decrease by 75%, ensuring smoothness of the motion.

[0143] The generation of the nonlinear damping field combines the damping adjustment parameter as a coefficient with the nonlinear damping matrix, and calculates an independent damping adjustment parameter for each characteristic direction, corresponding to the main diagonal elements of the damping matrix. In the precise operation task of the astronaut, if the eigenvalue of the first characteristic direction is 3.2 Nm / rad, the corresponding normalized stiffness value is 1.6, and the joint motion speed is 5 degrees / s, then the calculated damping adjustment parameter is 1.8 Nm·s / rad, indicating the damping strength in that direction.

[0144] The relative acceleration between adjacent joints reflects the coordination of joint motion. The angular acceleration is calculated by processing the joint angular velocity data through the differential algorithm, and then the relative acceleration between adjacent joints is obtained. When the astronaut performs a continuous change of direction, the relative acceleration between the shoulder joint and the elbow joint can reach 15 degrees / s 2 , indicating that the two joints are not synchronized and need to appropriately increase the damping to coordinate the motion.

[0145] The linear term coefficient of the damping adjustment parameter is affected by the relative acceleration and the stiffness deviation. The deviation between the coupling stiffness eigenvalue and the preset stiffness threshold is calculated, multiplied by the relative acceleration amplitude, and then multiplied by the proportional coefficient (typical value is 0.05 s / rad) to obtain the adjustment amount of the linear term coefficient. When the relative acceleration is large and the stiffness deviation is significant, the linear term coefficient increases, enhancing the damping effect. When the astronaut performs a high-load task, if the stiffness value of a certain characteristic direction is 4.0 Nm / rad, the preset threshold is 2.5 Nm / rad, and the relative acceleration is 20 degrees / s 2, the linear term coefficient increases by 0.15, which increases the damping strength by about 15%.

[0146] The update of the damping adjustment parameter comprehensively considers the relative acceleration and joint motion speed, and a two-dimensional mapping relationship is designed: when both the relative acceleration and the motion speed are small, the basic damping level is maintained; when the relative acceleration is large and the speed is small, the damping is significantly increased to suppress vibration; when the relative acceleration is small and the speed is large, the damping is reduced to avoid limiting normal motion; when both are large, the damping is moderately increased to balance stability and flexibility. When the astronaut is performing rapid attitude adjustment, if the relative acceleration is 25 degrees / second 2 and the joint speed is 35 degrees / second, the damping adjustment parameter is increased by 35%, which is sufficient to suppress vibration but will not significantly affect the smoothness of the action.

[0147] The updated damping adjustment parameter is substituted into the nonlinear damping field to realize dynamic damping control. The damping parameter is updated at a frequency of 200 Hz to ensure that the damping field can quickly respond to changes in motion state. In a microgravity environment, such high-frequency updates are crucial for suppressing vibration, as there is no gravitational damping and vibration decays slowly. Actual measurements show that the dynamic damping field can reduce the vibration decay time of the exoskeleton joint from 850 milliseconds in the traditional method to 210 milliseconds.

[0148] The application of damping torque follows the principle of energy dissipation. The damping torque to be applied to each joint is calculated according to the nonlinear damping field. The torque size is related to the joint angular velocity and the current damping coefficient, and the direction is always opposite to the motion direction. This design ensures that the damping force always does negative work, dissipating system energy. When the astronaut performs a hand swing action, the elbow joint angular velocity is 30 degrees / second, the current damping coefficient is 1.2 Nm·s / rad, and the applied damping torque is 0.63 Nm, opposite to the direction of elbow joint motion.

[0149] According to the distribution of the stiffness eigenvector, different intensities of damping are applied in different directions, preferentially dissipating excess energy in directions with larger stiffness. This directional dissipation mechanism allows the exoskeleton to maintain stability while not excessively limiting the astronaut's active motion. When the astronaut performs a compound action, the damping coefficient in the forward flexion direction of the shoulder joint is 1.5 Nm·s / rad, while in the internal and external rotation directions it is only 0.6 Nm·s / rad, allowing the shoulder joint to be stable in the forward flexion plane while maintaining flexibility in rotation.

[0150] In actual application scenarios, this technology significantly improves the performance of the astronaut exoskeleton. In a simulated space station extravehicular assembly task, after enabling the nonlinear damping field, the success rate of the astronaut completing a precise docking operation increased from 82% to 97%, and the average docking time was shortened by 38%. Astronaut feedback indicates that stability and responsiveness have improved significantly, especially in operations requiring high-precision control, where vibration suppression is most effective.

[0151] Adaptive adjustment of system parameters ensures that the exoskeleton adapts to different task requirements. In low-intensity fine operation, the damping strength is automatically increased to improve stability; in high-intensity exercise, the damping strength is reduced to enhance flexibility. This intelligent adjustment mechanism enables the exoskeleton to work efficiently in different environments inside and outside the space station. Test data shows that compared with the fixed parameter system, the adaptive damping field reduces the energy consumption of astronauts by 23%, increases the task completion accuracy by 42%, and significantly improves the work efficiency and comfort of astronauts in long-term space missions.

[0152] The exoskeleton motion strength adaptive adjustment and trajectory intelligent planning system of the embodiment of the application comprises:

[0153] The first unit is used for collecting motion data of astronauts in real time through sensors of an exoskeleton device, and generating a motion state feature vector;

[0154] The second unit is used for identifying feature phase points in the motion process based on the motion state feature vector, setting a torque control envelope according to the motion features of each feature phase point, dynamically calculating an angular momentum compensation value by analyzing the instantaneous change rate of joint angular acceleration and superimposing the angular momentum compensation value on the torque control envelope, and establishing a torque limit interval, and dynamically adjusting the upper and lower limits of the torque limit interval by real-time monitoring of joint range of motion, to ensure that the output torque is within a safe range;

[0155] The third unit is used for generating an expected motion trajectory based on the motion state feature vector, comparing the expected motion trajectory with an actual motion trajectory, and obtaining a trajectory deviation value;

[0156] The fourth unit is used for selecting a plurality of dynamic control points from the trajectory deviation value and performing spline interpolation to generate a deformable virtual channel to limit the motion range; calculating the inertia coupling effect in the joint motion process, and suppressing the interference of inertial force by adjusting the shape and stiffness characteristics of the deformable virtual channel, constructing a nonlinear damping field based on the joint coupling stiffness, and adaptively adjusting the damping adjustment parameters according to the relative motion acceleration between joints to dissipate excess inertial energy, to ensure that astronauts maintain a stable motion trajectory in a weightless environment.

[0157] In a third aspect, the embodiment of the application provides an electronic device, comprising:

[0158] a processor;

[0159] a memory for storing processor-executable instructions;

[0160] The processor is configured to call the instructions stored in the memory to execute the method described above.

[0161] In a fourth aspect, the present application provides a computer readable storage medium having stored thereon computer program instructions, which when executed by a processor, implement the method described above.

[0162] The present application can be a method, an apparatus, a system, and / or a computer program product. The computer program product can include a computer readable storage medium (or media) having computer readable program instructions thereon for performing various aspects of the present application.

[0163] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. Astronaut exoskeleton motion intensity self-adaptive adjustment and trajectory intelligent planning method, characterized in that, The method comprises the following steps: Real-time acquisition of astronaut motion data through sensors of an exoskeleton device to generate a motion state feature vector; Based on the motion state feature vector, feature phase points in the motion process are identified, and a torque control envelope is set according to the motion features of each feature phase point, including: obtaining the time domain features and frequency domain features of the motion state feature vector, and establishing a motion cycle feature map based on the time domain features and the frequency domain features; feature phase point identification is performed based on the motion cycle feature map, acceleration sudden change points are identified by calculating the angular acceleration change rate, angle extreme points are identified by analyzing the angular velocity zero-crossing points, and electromyographic signal peak points are identified by the sliding window method, and the acceleration sudden change points, the angle extreme points and the electromyographic signal peak points are combined to obtain a feature phase point set; for the feature phase points in the feature phase point set, the angular deviation of the joint angle vector corresponding to the feature phase point from the expected angle vector, and the angular velocity deviation of the angular velocity vector corresponding to the feature phase point from the expected angular velocity vector are calculated respectively, and the target torque value at the feature phase point is calculated based on the angular deviation and the angular velocity deviation; based on the motion features of each feature phase point, a weight coefficient of the feature phase point is set, the weight coefficient is combined with the target torque value, and the target torque value is interpolated by using a cubic spline interpolation method to obtain a torque control envelope; By analyzing the instantaneous change rate of joint angular acceleration, the angular momentum compensation value is dynamically calculated and superimposed on the torque control envelope, and a torque limiting interval is established, and the upper and lower limits of the torque limiting interval are dynamically adjusted by real-time monitoring of the joint range of motion, including: the instantaneous change rate of joint angular acceleration is calculated by using the central difference method, the time domain integral of the angular acceleration is calculated based on the instantaneous change rate to obtain angular momentum, the instantaneous change amount of the angular momentum and the deviation of the angular momentum from the expected angular momentum are combined to obtain an angular momentum compensation value, the angular momentum compensation value is superimposed on the torque output by the torque control envelope to obtain a superimposed torque; the ratio of the superimposed torque to the maximum bearing torque of the joint is calculated to obtain a torque utilization rate, a reference safety factor is set based on the torque utilization rate, and the superimposed torque and the reference safety factor are added and subtracted respectively to obtain the basic upper and lower limits of the torque limiting interval; a dynamic adjustment coefficient is calculated based on the change amount of the joint range of motion, the product of the dynamic adjustment coefficient and the upper and lower limit difference of the basic upper and lower limits is multiplied, and the product is subtracted from the upper limit of the basic upper and lower limits and superimposed on the lower limit of the basic upper and lower limits to obtain the dynamic upper and lower limits of the torque limiting interval; it is ensured that the output torque is within a safe range; Based on the motion state feature vector, an expected motion trajectory is generated, and the expected motion trajectory is compared with an actual motion trajectory to obtain a trajectory deviation value; The method comprises the following steps: selecting a plurality of dynamic control points from the trajectory deviation values and performing spline interpolation to generate a deformable virtual channel to limit the motion range, including: calculating the local curvature of the trajectory deviation values, selecting extreme points of the trajectory deviation values, extreme points of the local curvature and points with a speed change rate exceeding a preset speed threshold as initial dynamic control points; calculating the difference between the current joint activity and the reference activity to obtain a joint activity deviation value, calculating the difference between the current motion speed and the reference speed to obtain a speed deviation value, adding the joint activity deviation value and the speed deviation value to obtain a total adjustment amount, and adding the total adjustment amount to the position of the initial dynamic control points to obtain adjusted dynamic control points; performing cubic B-spline interpolation on the adjusted dynamic control points to obtain the center trajectory of the deformable virtual channel, calculating the width adjustment amount of the deformable virtual channel based on the joint activity deviation value and the speed deviation value, and expanding the width adjustment amount along both sides of the center trajectory to form the dynamic boundary of the deformable virtual channel; The method comprises the following steps: calculating the inertia coupling effect in the joint motion process, and suppressing the interference of inertial force by adjusting the shape and stiffness characteristics of the deformable virtual channel, including: obtaining the rotational inertia parameters and coupling inertia parameters of the joint and constructing a spatial inertia tensor matrix, obtaining the coupling inertia force based on the spatial inertia tensor matrix and the joint angular velocity vector; calculating the local deformation amount of the deformable virtual channel according to the amplitude of the coupling inertia force, adding the local deformation amount along the normal vector direction of the deformable virtual channel to obtain the deformed virtual channel; collecting the spatial posture information and angle posture information of the joint to obtain a joint posture matrix, taking the ratio of the amplitude of the coupling inertia force to the preset maximum allowed inertia force as a stiffness adjustment factor, multiplying the stiffness adjustment factor and the quadratic form of the joint posture matrix to obtain a stiffness increment matrix of the deformable virtual channel; calculating a compensation torque based on the coupling inertia force, adjusting the position and shape of the deformed virtual channel according to the compensation torque, and applying a constraint force to the deformed virtual channel based on the stiffness increment matrix to realize dynamic suppression of the inertia force interference; A nonlinear damping field based on joint coupling stiffness is constructed, and the damping adjustment parameter is adaptively adjusted according to the relative motion acceleration between joints to dissipate excess inertial energy, so as to ensure that the astronaut maintains a stable motion trajectory in a weightless environment.

2. The method of claim 1, wherein, The method comprises the following steps: Obtaining a joint configuration vector, calculating a generalized stiffness matrix based on the joint configuration vector, and performing eigenvalue decomposition on the generalized stiffness matrix to obtain a coupling stiffness eigenvector and a coupling stiffness eigenvalue; Based on the coupling stiffness eigenvector, a nonlinear damping matrix is established, a damping adjustment parameter is constructed according to the coupling stiffness eigenvalue and an exponential decay term of the joint motion speed, the damping adjustment parameter is set as the coefficient of the nonlinear damping matrix, and a nonlinear damping field is obtained. The relative acceleration between adjacent joints is calculated, the linear term coefficient of the damping adjustment parameter is updated according to the deviation amount of the amplitude of the relative acceleration from the preset stiffness threshold value relative to the coupling stiffness characteristic value, the exponential decay coefficient of the damping adjustment parameter is updated according to the product of the amplitude of the relative acceleration and the square of the joint movement speed, the updated damping adjustment parameter is substituted into the nonlinear damping field, and the directional dissipation of the redundant inertial energy in the movement process is realized by using the damping torque exerted by the nonlinear damping field.

3. A system for astronaut exoskeleton motion intensity self-adaptive adjustment and trajectory intelligent planning, for implementing the method according to any one of claims 1-2, characterized in that, The method comprises the following steps: A first unit is configured to collect astronaut movement data in real time through sensors of an exoskeleton device and generate a movement state feature vector; A second unit is configured to identify feature phase points in the movement process based on the movement state feature vector, set a torque control envelope according to movement features of each feature phase point, dynamically calculate an angular momentum compensation value by analyzing the instantaneous change rate of joint angular acceleration and superimpose the angular momentum compensation value on the torque control envelope, and establish a torque limit interval, dynamically adjust the upper and lower limits of the torque limit interval by monitoring joint range of motion in real time, and ensure that the output torque is within a safe range; A third unit is configured to generate an expected movement trajectory based on the movement state feature vector, compare the expected movement trajectory with an actual movement trajectory, and obtain a trajectory deviation value; A fourth unit is configured to select a plurality of dynamic control points from the trajectory deviation value and perform spline interpolation to generate a deformable virtual channel to limit the movement range; The inertia coupling effect in the joint movement process is calculated, the interference of inertial force is suppressed by adjusting the shape and stiffness characteristics of the deformable virtual channel, a nonlinear damping field based on joint coupling stiffness is constructed, the damping adjustment parameter is adaptively adjusted according to the relative movement acceleration between joints to dissipate redundant inertial energy, and the astronaut can maintain a stable movement trajectory in a weightless environment.

4. An electronic device, comprising: The method comprises the following steps: A processor; A memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the method of any one of claims 1 to 2.

5. A computer-readable storage medium having stored thereon computer program instructions, wherein, The computer program instructions are executed by the processor to implement the method of any one of claims 1 to 2.

Images