Intelligent guidance feedback method and system for astronaut physical exercise process

Abstract

The present application provides an astronaut physical exercise process intelligent guidance feedback method and system, which relates to the technical field of data feedback, comprising the following steps: obtaining three-dimensional motion trajectory data in the training process of astronauts, performing kinematic analysis calculation to obtain joint angle sequence and limb space posture parameters, comparing and analyzing the deviation vector with a standard action template, generating posture correction instructions and real-time feedback, and simultaneously regulating the load parameters of the training equipment, thereby realizing precise guidance and intelligent correction in the training process of astronauts, and improving the training effect and safety.

Description

Technical Field

[0001] This invention relates to the field of data feedback technology, and in particular to an intelligent guidance and feedback method and system for astronaut physical training. Background Technology

[0002] With the rapid development of aerospace technology and the increasing number of space exploration missions, astronaut physical training in orbit has become a crucial aspect of ensuring the health of astronauts during long-term space missions, attracting increasing attention. Prolonged exposure to the microgravity environment of space can lead to a series of physiological problems for astronauts, including muscle atrophy, decreased bone density, and cardiovascular dysfunction. Therefore, regular physical exercise is necessary to maintain bodily functions. Current astronaut physical training primarily relies on specialized training equipment within the space station, such as treadmills, cycling machines, and resistance training devices. However, with the advancement of space station technology and the extension of space mission durations, the demand for precision, scientific rigor, and personalization in astronaut physical training is also increasing.

[0003] Traditional physical training methods lack real-time, precise motion capture and evaluation mechanisms, making it difficult to comprehensively and accurately monitor astronauts' training movements. This leads to inaccurate assessments of training effectiveness and may result in incorrect postures and potential injury risks during training. Existing training systems lack intelligent movement guidance capabilities, failing to provide personalized training suggestions and real-time adjustments based on individual astronaut differences and specific space mission requirements, thus limiting the relevance and effectiveness of training. Insufficient coordination between traditional training equipment and feedback systems hinders the intelligent linkage between training load and movement correction, resulting in imprecise load adjustment during training and an inability to dynamically optimize training parameters based on astronauts' real-time performance, ultimately impacting overall training effectiveness.

[0004] With the development of artificial intelligence, computer vision, and intelligent human-computer interaction technologies, applying these advanced technologies to the physical training of astronauts and building an intelligent and precise training guidance and feedback system is of great significance for improving the quality and efficiency of astronaut training in space, and is also an inevitable trend in the future development of astronaut training systems. Summary of the Invention

[0005] The present invention provides an intelligent guidance and feedback method and system for astronaut physical training, which can solve the problems in the prior art.

[0006] A first aspect of the present invention provides an intelligent guidance and feedback method for astronaut physical training, comprising:

[0007] Real-time three-dimensional motion trajectory data of astronauts at multiple joint points during training is acquired using a space-distributed motion capture device.

[0008] The multi-joint three-dimensional motion trajectory data is spatially clustered according to human anatomy partitioning rules to obtain multiple joint set sets. Multiple kinematic skeletal chains are constructed according to hierarchical parent-child relationships. Inverse kinematic analysis is performed sequentially from the terminal joints along the skeletal chains toward the root joints to obtain the joint angle sequence and limb spatial posture parameters of the astronaut's current training movements.

[0009] Based on the joint angle sequence and the limb spatial posture parameters, the astronaut's current training movements are aligned with the preset space mission standard movement template in a multi-dimensional space to obtain the posture deviation vector;

[0010] Based on the posture deviation vector, calculate the target joint angle or target limb spatial position that the deviation dimension needs to achieve, plan the optimal adjustment path from the current posture state to the target joint angle or target limb spatial position, and generate posture correction instructions that include step-by-step execution instructions and execution timing control;

[0011] The system outputs attitude correction commands to the astronauts through a real-time feedback device and simultaneously adjusts the drag load parameters of the training equipment.

[0012] The multi-joint 3D motion trajectory data is spatially clustered according to human anatomy zoning rules to obtain multiple joint set sets. Multiple kinematic skeletal chains are constructed according to hierarchical parent-child relationships. Inverse kinematic analysis is performed sequentially from the terminal joints along the skeletal chains towards the root joints to obtain the joint angle sequence and limb spatial posture parameters of the astronaut's current training movements, including:

[0013] The multi-joint 3D motion trajectory data is spatially clustered according to human anatomical zoning rules to obtain multiple joint set sets;

[0014] Based on predefined skeletal connection relationships, multiple kinematic skeleton chains are constructed from the joints in the multiple sets of joints according to hierarchical parent-child relationships. Each kinematic skeleton chain contains a root joint and an end joint.

[0015] For each kinematic skeletal chain, inverse kinematic analysis is performed sequentially from the terminal joint along the skeletal chain towards the root joint. Based on the three-dimensional spatial positional relationship between adjacent joints and the bone length constraint, the three degrees of freedom rotation angles of each joint are solved to obtain the joint angle sequence.

[0016] Based on the angle values ​​of the root joints in the joint angle sequence and the overall spatial orientation of the kinematic skeletal chain, the tilt angle of the astronaut's torso relative to the direction of gravity and the coordinates of the center of mass are calculated to obtain the spatial posture parameters of the limbs.

[0017] For each kinematic skeletal chain, inverse kinematic analysis is performed sequentially from the terminal joint along the skeletal chain towards the root joint. Based on the three-dimensional spatial positional relationship between adjacent joints and the bone length constraint, the three degrees of freedom rotation angles of each joint are solved, including:

[0018] Starting from the terminal joint, each joint is traversed sequentially along the kinematic skeleton chain toward the root joint. For the currently traversed joint and its parent joint, a connection vector is calculated based on the three-dimensional spatial positional relationship between them. The connection vector is then normalized in combination with the bone length constraint to obtain a unit direction vector sequence.

[0019] For each unit direction vector in the unit direction vector sequence, the unit direction vector is decomposed into three orthogonal rotation components relative to the local coordinate system of the parent joint. By calculating the angles between the three orthogonal rotation components and each coordinate axis of the local coordinate system of the parent joint, the rotation angles of each joint in the pitch, yaw, and roll directions are solved to obtain the joint angle sequence.

[0020] The three-degree-of-freedom rotation angle of each joint in the joint angle sequence is compared one by one with the preset joint physiological range of motion constraints. When a rotation angle that exceeds the joint physiological range of motion constraints is detected, an angle correction mechanism is triggered.

[0021] For rotation angles that exceed the physiological range of motion constraints of the joint, the maximum allowable rotation angle of the joint is calculated based on the spacesuit's limiting constraints, and the excess rotation angle is redistributed to the joints adjacent to the joint in the kinematic skeletal chain, thus updating the joint angle sequence.

[0022] Based on the joint angle sequence and the limb spatial posture parameters, the astronaut's current training movements are aligned with the preset space mission standard movement template in a multi-dimensional space, resulting in a posture deviation vector including:

[0023] Align the joint angle sequence with the standard joint angle sequence in the standard action template for aerospace missions along the time axis to establish a temporal mapping relationship between the current training action and the standard action template for aerospace missions.

[0024] Based on the aforementioned temporal mapping relationship, for each joint in the joint angle sequence, the angle difference between its three-degree-of-freedom rotation angle and the corresponding rotation angle of the corresponding joint at the corresponding time in the standard joint angle sequence is calculated to obtain the joint angle deviation matrix;

[0025] For the trunk tilt angle and the center of mass position coordinates in the limb spatial posture parameters, calculate their spatial differences with the corresponding parameters in the standard limb spatial posture parameters to obtain the posture spatial deviation vector;

[0026] The joint angle deviation matrix and the attitude space deviation vector are weighted and fused according to a preset dimension weight coefficient to generate the attitude deviation vector.

[0027] Based on the posture deviation vector, calculate the target joint angle or target limb spatial position that the deviation dimension needs to achieve, plan the optimal adjustment path from the current posture state to the target joint angle or target limb spatial position, and generate posture correction instructions that include step-by-step execution instructions and execution timing control, including:

[0028] Based on the joint angle deviation value or limb spatial position deviation value corresponding to the deviation dimension in the posture deviation vector, and combined with the standard joint angle sequence and standard limb spatial posture parameters in the standard action template for aerospace missions, the target joint angle or target limb spatial position that the deviation dimension needs to achieve is calculated, and a graded correction target set is generated.

[0029] Based on each correction target in the graded correction target set, the tolerable adjustment rate of the joint or limb is determined according to the heart rate index in the astronaut's current physiological state parameters, and the optimal adjustment path from the current posture state to the target joint angle or the target limb spatial position is planned;

[0030] The intermediate posture node sequence in the optimal adjustment path is converted into specific joint rotation direction indicators, joint rotation amplitude values, and limb movement direction vectors, generating posture correction instructions that include step-by-step execution instructions and execution timing control.

[0031] Based on the heart rate index in the astronaut's current physiological parameters, the tolerable adjustment rate of the joint or limb is determined, and the optimal adjustment path from the current posture to the target joint angle or the target limb spatial position is planned, including:

[0032] Based on heart rate indicators, the cardiovascular load state is determined to belong to the load zone, and a corresponding rate adjustment coefficient is assigned to the adjustment movement of the joint or limb. The rate adjustment coefficient is used to regulate the execution speed of the joint or limb adjusting from the current posture state to the target joint angle or the target limb spatial position.

[0033] Calculate the total angular change or total spatial displacement of the joint or limb from its current posture to the target joint angle or the target limb's spatial position. Divide the total angular change or total spatial displacement into segments according to the rate adjustment coefficient to generate multiple sub-adjustment angles or sub-adjustment displacements.

[0034] The endpoint of each sub-adjustment angle or sub-adjustment displacement is taken as an intermediate posture node. The intermediate posture nodes are sequentially set on the adjustment trajectory from the current posture state to the target joint angle or the target limb spatial position to form an intermediate posture node sequence.

[0035] For the sub-adjustment angle or sub-adjustment displacement between two adjacent intermediate attitude nodes in the intermediate attitude node sequence, the execution time between nodes required to complete the sub-adjustment angle or sub-adjustment displacement is calculated in conjunction with the rate adjustment coefficient;

[0036] The intermediate posture node sequence is combined with the execution time between each node to generate the optimal adjustment path from the current posture state to the target joint angle or the target limb spatial position.

[0037] A second aspect of the present invention provides an intelligent guidance and feedback system for astronaut physical training, comprising:

[0038] The first unit is used to acquire real-time three-dimensional motion trajectory data of astronauts at multiple joints during training through a space-distributed motion capture device.

[0039] The second unit is used to perform spatial clustering on the multi-joint three-dimensional motion trajectory data according to human anatomy partitioning rules to obtain multiple joint set, construct multiple kinematic skeletal chains according to hierarchical parent-child relationships, and perform inverse kinematic analysis calculations sequentially from the terminal joints along the skeletal chains to the root joints to obtain the joint angle sequence and limb spatial posture parameters of the astronaut's current training movements;

[0040] The third unit is used to perform multi-dimensional spatial alignment between the astronaut's current training movements and the preset space mission standard movement template based on the joint angle sequence and the limb spatial posture parameters, so as to obtain the posture deviation vector;

[0041] The fourth unit is used to calculate the target joint angle or target limb spatial position that the deviation dimension needs to reach based on the posture deviation vector, plan the optimal adjustment path from the current posture state to the target joint angle or target limb spatial position, and generate posture correction instructions that include step-by-step execution instructions and execution timing control;

[0042] The fifth unit is used to output attitude correction commands to astronauts through a real-time feedback device and to simultaneously adjust the drag load parameters of the training equipment.

[0043] A third aspect of the present invention provides an electronic device, comprising:

[0044] processor;

[0045] Memory used to store processor-executable instructions;

[0046] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0047] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0048] The beneficial effects of this application are as follows:

[0049] This method enables precise capture and analysis of astronaut training movements. By constructing multiple kinematic skeletal chains and performing inverse analytical calculations, it can accurately obtain joint angle sequences and limb spatial posture parameters, providing a reliable data foundation for subsequent movement evaluation and correction.

[0050] This invention aligns the astronaut's current training movements with a preset standard movement template for space missions in a multi-dimensional space, accurately calculates the attitude deviation vector, and plans the optimal adjustment path accordingly, making attitude correction more scientific and reasonable and effectively improving training quality.

[0051] This method outputs attitude correction commands to astronauts through a real-time feedback device and simultaneously adjusts the resistance load parameters of the training equipment, realizing closed-loop control of the training process. This not only improves the accuracy and safety of astronaut physical training, but also allows for dynamic adjustment of training intensity based on actual conditions, resulting in better training effects. Attached Figure Description

[0052] Figure 1 This is a flowchart illustrating the intelligent guidance and feedback method for astronaut physical training according to an embodiment of the present invention.

[0053] Figure 2 This is a flowchart of inverse kinematics calculation for processing multi-joint three-dimensional motion trajectory data in an embodiment of the present invention. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0056] Figure 1 This is a flowchart illustrating the intelligent guidance and feedback method for astronaut physical training according to an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes:

[0057] Real-time three-dimensional motion trajectory data of astronauts at multiple joint points during training is acquired using a space-distributed motion capture device.

[0058] The multi-joint three-dimensional motion trajectory data is spatially clustered according to human anatomy partitioning rules to obtain multiple joint set sets. Multiple kinematic skeletal chains are constructed according to hierarchical parent-child relationships. Inverse kinematic analysis is performed sequentially from the terminal joints along the skeletal chains toward the root joints to obtain the joint angle sequence and limb spatial posture parameters of the astronaut's current training movements.

[0059] Based on the joint angle sequence and the limb spatial posture parameters, the astronaut's current training movements are aligned with the preset space mission standard movement template in a multi-dimensional space to obtain the posture deviation vector;

[0060] Based on the posture deviation vector, calculate the target joint angle or target limb spatial position that the deviation dimension needs to achieve, plan the optimal adjustment path from the current posture state to the target joint angle or target limb spatial position, and generate posture correction instructions that include step-by-step execution instructions and execution timing control;

[0061] The system outputs attitude correction commands to the astronauts through a real-time feedback device and simultaneously adjusts the drag load parameters of the training equipment.

[0062] In one optional implementation, the multi-joint three-dimensional motion trajectory data is spatially clustered according to human anatomical partitioning rules to obtain multiple joint set sets. Multiple kinematic skeletal chains are constructed according to hierarchical parent-child relationships. Inverse kinematic analysis is performed sequentially from the terminal joints along the skeletal chains towards the root joints to obtain the joint angle sequence and limb spatial posture parameters of the astronaut's current training movements, including:

[0063] The multi-joint 3D motion trajectory data is spatially clustered according to human anatomical zoning rules to obtain multiple joint set sets;

[0064] Based on predefined skeletal connection relationships, multiple kinematic skeleton chains are constructed from the joints in the multiple sets of joints according to hierarchical parent-child relationships. Each kinematic skeleton chain contains a root joint and an end joint.

[0065] For each kinematic skeletal chain, inverse kinematic analysis is performed sequentially from the terminal joint along the skeletal chain towards the root joint. Based on the three-dimensional spatial positional relationship between adjacent joints and the bone length constraint, the three degrees of freedom rotation angles of each joint are solved to obtain the joint angle sequence.

[0066] Based on the angle values ​​of the root joints in the joint angle sequence and the overall spatial orientation of the kinematic skeletal chain, the tilt angle of the astronaut's torso relative to the direction of gravity and the coordinates of the center of mass are calculated to obtain the spatial posture parameters of the limbs.

[0067] like Figure 2 As shown, the method includes:

[0068] The system acquires three-dimensional motion trajectory data of multiple joints during astronaut training, which can be obtained through a motion capture system. The system uses infrared reflective markers or inertial sensors to collect three-dimensional spatial coordinate data of the astronaut's main joints, including the head, shoulders, elbows, wrists, hips, knees, and ankles. For example, for a complete squat training movement, the system captures the motion trajectory of each joint in three-dimensional space throughout the entire process of the astronaut going from standing to squatting and back to standing.

[0069] The acquired multi-joint 3D motion trajectory data was spatially clustered according to human anatomical partitioning rules. Based on human anatomical structure, the joints were divided into multiple functional regions, such as the head region, upper limb region, trunk region, and lower limb region. For example, the head region includes the vertex, forehead, and mandible; the upper limb region includes the shoulder, elbow, and wrist joints; the trunk region includes the cervical, thoracic, and lumbar vertebrae; and the lower limb region includes the hip, knee, and ankle joints. The K-means clustering algorithm was used to cluster the joints into different sets based on their 3D spatial distribution characteristics. The clustering process considered the spatial distance and functional correlation between joints to ensure that joints within the same functional region were correctly clustered. In practice, the system clustered 20 joint data points of an astronaut into four main joint sets, corresponding to the head, upper limbs, trunk, and lower limbs, respectively.

[0070] Based on predefined skeletal connections, multiple kinematic skeletal chains are constructed from joints in a hierarchical parent-child relationship. The human skeletal structure has a clear hierarchical parent-child relationship; for example, in the upper limb skeletal chain, the shoulder joint is the root joint, the elbow joint is the intermediate joint, and the wrist joint is the terminal joint. Based on this anatomical relationship, the system constructs skeletal chains including the head chain, left upper limb chain, right upper limb chain, trunk chain, left lower limb chain, and right lower limb chain. Each skeletal chain contains clearly defined root and terminal joints, and establishes connections between these joints. For example, the root joint of the right upper limb skeletal chain is the right shoulder joint, the intermediate joint is the right elbow joint, and the terminal joint is the right wrist joint; the bone length constraints are an upper arm length of 35 cm and a forearm length of 30 cm.

[0071] For each kinematic skeletal chain, inverse kinematic analysis is performed sequentially from the distal joints along the chain towards the root joints. Taking the right upper limb skeletal chain as an example, the three-dimensional spatial coordinates of the right wrist joint (the distal joint) are first determined. Then, based on the length constraints of the right forearm bones, the three-degree-of-freedom rotation angles of the right elbow joint are calculated. The calculation method derives the flexion-extension angle, internal and external rotation angle, and lateral bending angle of the right elbow joint from the position vector between the right wrist and right elbow joints. Similarly, based on the length constraints of the right upper arm bones, the three-degree-of-freedom rotation angles of the right shoulder joint are calculated. In actual calculations, when an astronaut completes a push-up, the flexion-extension angle of the right elbow joint changes from an initial 175 degrees to a minimum of 80 degrees, and then returns to 175 degrees, exhibiting a clear periodic variation.

[0072] After performing inverse kinematic analysis on all skeletal chains, a complete sequence of joint angles was obtained. These angle values ​​describe the rotational state of each joint during astronaut training and can be used to assess the standardization of movements. For example, in a standard squat, the flexion and extension angles of the knee joint should vary between 40 and 160 degrees, and the difference in angle between the left and right knee joints should not exceed 10 degrees to ensure balance.

[0073] Based on the angle values ​​of the root joints in the joint angle sequence and the overall spatial orientation of the kinematic skeletal chain, the tilt angle of the astronaut's torso relative to the direction of gravity and the coordinates of the center of mass are calculated. The torso tilt angle is calculated using the three-dimensional rotation angle of the root joint of the torso skeletal chain (usually the center point of the hip joint). The coordinates of the center of mass are calculated based on the mass distribution and spatial position of each bone segment. For example, in a forward flexion training movement, the system detected that the forward tilt angle of the astronaut's torso relative to the direction of gravity increased from an initial 5 degrees to a maximum of 85 degrees, while the lateral tilt angle remained within 3 degrees, indicating good planarity of the movement. The center of mass position decreased from an initial height of 0.9 meters to 0.7 meters, a horizontal displacement of 0.2 meters, reflecting the characteristics of forward shift and sinking of the center of gravity.

[0074] In one optional implementation, for each kinematic skeletal chain, inverse kinematic analysis is performed sequentially from the terminal joint along the skeletal chain towards the root joint. Based on the three-dimensional spatial positional relationship between adjacent joints and the bone length constraint, the three degrees of freedom rotation angles of each joint are solved, including:

[0075] Starting from the terminal joint, each joint is traversed sequentially along the kinematic skeleton chain toward the root joint. For the currently traversed joint and its parent joint, a connection vector is calculated based on the three-dimensional spatial positional relationship between them. The connection vector is then normalized in combination with the bone length constraint to obtain a unit direction vector sequence.

[0076] For each unit direction vector in the unit direction vector sequence, the unit direction vector is decomposed into three orthogonal rotation components relative to the local coordinate system of the parent joint. By calculating the angles between the three orthogonal rotation components and each coordinate axis of the local coordinate system of the parent joint, the rotation angles of each joint in the pitch, yaw, and roll directions are solved to obtain the joint angle sequence.

[0077] The three-degree-of-freedom rotation angle of each joint in the joint angle sequence is compared one by one with the preset joint physiological range of motion constraints. When a rotation angle that exceeds the joint physiological range of motion constraints is detected, an angle correction mechanism is triggered.

[0078] For rotation angles that exceed the physiological range of motion constraints of the joint, the maximum allowable rotation angle of the joint is calculated based on the spacesuit's limiting constraints, and the excess rotation angle is redistributed to the joints adjacent to the joint in the kinematic skeletal chain, thus updating the joint angle sequence.

[0079] The inverse kinematic analysis (IMA) calculation begins at the terminal joints of each kinematic skeletal chain, employing a depth-first traversal algorithm to sequentially visit each joint along the chain towards the root joint. During traversal, a spatial geometric relationship is established between the currently processed joint and its parent joint. The system obtains the connection vector by calculating the difference in the three-dimensional coordinates of the two joints. The connection vector is calculated using a Cartesian coordinate system, where the X-axis represents the anterior-posterior direction, the Y-axis represents the lateral direction, and the Z-axis represents the vertical direction. When the current joint coordinates are (x1, y1, z1) and the parent joint coordinates are (x2, y2, z2), the connection vector is (x1-x2, y1-y2, z1-z2). The system also maintains a skeletal length constraint database, which stores the standard length ranges for each bone segment, including anatomical reference values ​​such as the adult male humerus length of 280-360 mm, forearm length of 240-280 mm, thigh length of 400-480 mm, and lower leg length of 350-430 mm.

[0080] The length normalization of the connection vector is achieved by calculating the Euclidean norm of the vector. The vector length is obtained by taking the square root of the sum of the squares of each component of the connection vector. The normalization process divides each component of the connection vector by the vector length, generating a unit direction vector with a modulus of 1. When the calculated vector length deviates from the preset skeletal length constraint by more than 5%, the system triggers a length correction mechanism to adjust the vector length to the nearest value within the constraint range, ensuring biomechanical rationality. The unit direction vector sequence is stored sequentially according to the hierarchical structure of the skeletal chain. Each vector contains three components and satisfies the constraint that the sum of their squares equals 1. The system uses 32-bit floating-point numbers to store the vector components, achieving a precision of 6 decimal places, which meets the accuracy requirements of human motion analysis.

[0081] The decomposition process of the unit direction vector is based on the local coordinate system of each joint. The origin of the local coordinate system is located at the parent joint, and the three coordinate axes correspond to the standard directions of human anatomy. The system obtains the orthogonal rotation components by calculating the projection of the unit direction vector onto the three axes of the local coordinate system. The projection calculation uses vector dot product operations. The pitch rotation component corresponds to the rotation about the X-axis and is obtained by calculating the angle between the projection of the unit direction vector onto the YZ plane and the positive Z-axis. The yaw rotation component corresponds to the rotation about the Z-axis and is obtained by calculating the angle between the projection of the unit direction vector onto the XY plane and the positive X-axis. The roll rotation component corresponds to the rotation about the Y-axis and is obtained by calculating the angle between the projection of the unit direction vector onto the XZ plane and the positive X-axis.

[0082] The angle calculation uses the inverse cosine function, with the input parameter being the dot product of two unit vectors, and the output angle range is 0 to 180 degrees. The system converts the calculation result into degrees and determines the sign of the angle based on the quadrant information of the vectors, ensuring that the angle value can completely represent the rotational state in three-dimensional space. The joint angle sequence stores the rotation angle of each joint in the form of triples (pitch angle, yaw angle, roll angle), with an angle accuracy of 0.1 degrees. The system also records the confidence score for each joint, which is a comprehensive evaluation based on the quality of the input data, the stability of the calculation process, and the biomechanical rationality of the result, with a value ranging from 0 to 1.

[0083] The joint physiological range of motion constraint database contains the movement limit parameters of major human joints. The shoulder joint has a pitch range of -60 to 180 degrees, a yaw range of -90 to 90 degrees, and a roll range of -90 to 90 degrees. The elbow joint primarily restricts pitch movement, with a range of 0 to 150 degrees; its yaw and roll ranges are smaller, at -15 to 15 degrees respectively. The hip joint has a pitch range of -30 to 120 degrees, a yaw range of -45 to 45 degrees, and a roll range of -30 to 30 degrees. The knee joint primarily moves in the pitch direction, with a range of 0 to 140 degrees. The ankle joint has a pitch range of -30 to 45 degrees and a yaw range of -25 to 25 degrees. The system compares the calculated joint angles with the corresponding physiological range of motion constraints one by one. When any angle component exceeds the constraint range, the joint is immediately marked and the extent of the exceedance is recorded.

[0084] The angle correction mechanism is triggered by factors including angles exceeding the physiological range, abnormal rates of angle change, and abnormal coordination between adjacent joints. The system maintains spacesuit constraint parameters, determined based on the spacesuit's material properties, joint design, and sealing requirements. The spacesuit imposes stricter constraints on the shoulder joint than the physiological range, reducing the pitch range to -40 to 160 degrees, yaw to -70 to 70 degrees, and roll to -70 to 70 degrees. Under spacesuit constraints, the elbow joint has a pitch range of 10 to 130 degrees, the hip joint a pitch range of -20 to 100 degrees, and the knee joint a pitch range of 10 to 120 degrees. The system recalculates the maximum permissible rotation angle for each joint based on the spacesuit constraints, and when an excess is detected, it redistributes the excess to adjacent joints according to a preset weighting.

[0085] The angle redistribution strategy prioritizes the continuity and biomechanical rationality of the kinetic chain, maintaining a coupling weight matrix between joints that describes the motion coordination between different joints. The coupling weight between the shoulder and elbow joints is 0.7, indicating that 70% of the excess angle can be transferred to the elbow joint when the shoulder joint is constrained. The coupling weight between the hip and knee joints is 0.8, and the coupling weight between the segments of the spine is 0.6. The redistribution process uses an iterative algorithm. In each iteration, the excess angle is allocated to adjacent joints according to the coupling weight, and then it is checked whether the receiving joint also exceeds the constraint. If it does, the transfer continues to the next level joint. The iteration process executes a maximum of 5 rounds. If there are still excess angles that cannot be allocated, the system allocates these angles to the root joint of the kinetic chain, usually the trunk or pelvis.

[0086] The updated joint angle sequence maintains the original data structure and storage format, while updating the correction flag and correction magnitude record for each joint. The correction flag uses 8-bit binary encoding to represent the correction status in pitch, yaw, and roll directions, as well as the correction source information. The correction magnitude record includes the original angle, the angle before correction, the angle after correction, and the angle amount allocated to other joints. This information is used for subsequent motion quality assessment and feedback optimization. The system uses a circular buffer to store the joint angle sequence of the most recent 100 frames, supporting real-time processing and historical data backtracking analysis.

[0087] In a real-world application case, an astronaut was performing upper limb extension training. The initial angles of the right shoulder joint were collected as follows: pitch 200 degrees, yaw 100 degrees, and roll -100 degrees, all exceeding the spacesuit's limits. The system calculated that the pitch angle exceeded the limit by 40 degrees, the yaw angle by 30 degrees, and the roll angle by 30 degrees. Based on the coupling weight matrix, 28 degrees (40 × 0.7) of pitch, 21 degrees (30 × 0.7) of yaw, and 21 degrees (30 × 0.7) of roll were allocated to the elbow joint. After receiving the allocated angles, the elbow joint had a pitch angle of 35 degrees, a yaw angle of 8 degrees, and a roll angle of 6 degrees, all within the allowable range. The remaining 12 degrees of pitch, 9 degrees of yaw, and 9 degrees of roll were allocated to the trunk joints. The final updated joint angle sequence is as follows: right shoulder joint pitch 160 degrees, yaw 70 degrees, roll -70 degrees; right elbow joint pitch 35 degrees, yaw 8 degrees, roll 6 degrees; trunk joint pitch 12 degrees, yaw 9 degrees, roll 9 degrees. All angles meet the constraints and maintain the biomechanical rationality of the movement.

[0088] In one optional implementation, based on the joint angle sequence and the limb spatial posture parameters, the astronaut's current training movements are aligned with a preset space mission standard movement template in a multi-dimensional spatial manner to obtain a posture deviation vector, including:

[0089] Align the joint angle sequence with the standard joint angle sequence in the standard action template for aerospace missions along the time axis to establish a temporal mapping relationship between the current training action and the standard action template for aerospace missions.

[0090] Based on the aforementioned temporal mapping relationship, for each joint in the joint angle sequence, the angle difference between its three-degree-of-freedom rotation angle and the corresponding rotation angle of the corresponding joint at the corresponding time in the standard joint angle sequence is calculated to obtain the joint angle deviation matrix;

[0091] For the trunk tilt angle and the center of mass position coordinates in the limb spatial posture parameters, calculate their spatial differences with the corresponding parameters in the standard limb spatial posture parameters to obtain the posture spatial deviation vector;

[0092] The joint angle deviation matrix and the attitude space deviation vector are weighted and fused according to a preset dimension weight coefficient to generate the attitude deviation vector.

[0093] When astronauts align their current training movements with the preset standard movement template for space missions in a multi-dimensional spatial manner, the joint angle sequence must first be aligned with the standard joint angle sequence in terms of time axis. Time axis alignment is achieved through a dynamic time warping algorithm. This algorithm constructs a cost matrix to measure the similarity between each time frame in the current training movement sequence and each time frame in the standard movement template. Each element of the cost matrix represents the cumulative Euclidean distance between all joint angles between two time frames. Specifically, the inter-frame distance is obtained by taking the square root of the sum of the squared differences between the three-degree-of-freedom rotation angles of all joints in the i-th frame of the current training movement sequence and the j-th frame of the standard movement template. This distance is used as the element value in the i-th row and j-th column of the cost matrix.

[0094] The Dynamic Time Warping algorithm starts from the top left corner of the cost matrix and searches towards the bottom right corner according to the principle of minimizing the cumulative cost path. During the path search, each step can choose to move one grid to the right (corresponding to advancing one frame of the standard template time), move one grid down (corresponding to advancing one frame of the current action time), or move one grid diagonally (corresponding to advancing both simultaneously by one frame). The path selection criterion is to minimize the cumulative cost. The cumulative cost is calculated by adding the cost of the current cell to the minimum cumulative cost among the adjacent predecessor cells. The adjacent predecessor cells include the left cell, the bottom cell, and the bottom-left diagonal cell.

[0095] The search process employs a dynamic programming strategy to fill the cumulative cost matrix row by row or column by column from the starting point. After reaching the endpoint, the path is backtracked to obtain the optimal temporal mapping relationship. This mapping relationship maps each time frame of the current training action sequence to the corresponding time frame of the standard action template, establishing a one-to-one or one-to-many temporal alignment relationship to ensure accurate correspondence in the time dimension for subsequent bias calculations.

[0096] After establishing the temporal mapping relationship, the angle difference is calculated for the three-degree-of-freedom rotation angles of each joint in the joint angle sequence. The three-degree-of-freedom rotation angles include pitch, yaw, and roll, representing the amount of rotation of the joint around the horizontal, vertical, and forward / backward axes, respectively, and Euler angles are used to describe the joint's spatial attitude. Based on the temporal mapping relationship, for the i-th frame in the current training action sequence, the corresponding frame index j mapped to it in the standard action template is found, and the three-degree-of-freedom rotation angle values ​​of all joints in the current frame are extracted and compared with the three-degree-of-freedom rotation angle values ​​of the corresponding joints in the j-th frame of the standard template.

[0097] For each rotational degree of freedom of each joint, calculate the angle difference. The angle difference equals the angle value of that joint and that degree of freedom in the current training movement minus the angle value of the corresponding joint and that degree of freedom in the standard template. Since rotational angles are periodic and their effective range is typically defined between -180° and +180°, the angle difference calculation must handle cases where the angle crosses this boundary. When the absolute value of the calculated original difference is greater than 180°, normalize the difference. If the difference is greater than 180°, subtract 360° from the difference; if the difference is less than -180°, add 360° to the difference. This ensures that the normalized difference is strictly controlled within the range of -180° to +180°, avoiding calculation errors caused by the periodicity of angle representation.

[0098] The three-degree-of-freedom angle differences of all joints across all time frames are organized into a three-dimensional tensor structure according to the joint index, time frame index, and degree-of-freedom index. This tensor is the joint angle deviation matrix. The first dimension of the joint angle deviation matrix corresponds to the number of joints, which is usually eighteen, including the bilateral shoulder, elbow, wrist, hip, knee, and ankle joints, as well as the trunk and neck joints. The second dimension corresponds to the number of time frames, which depends on the length of the motion sequence. The third dimension is fixed at three, corresponding to the three degrees of freedom of pitch, yaw, and roll. Each element in the matrix is ​​a specific angle difference in degrees.

[0099] The spatial difference between the trunk tilt angle and the center of mass coordinates in the limb spatial posture parameters is calculated. The trunk tilt angle includes the forward / backward tilt angle and the left / right lateral tilt angle relative to the direction of gravity. The forward / backward tilt angle is calculated by the angle between the projection of the trunk axis vector onto the sagittal plane and the vertical direction. The trunk axis vector is defined as a three-dimensional vector pointing from the cervical vertebrae to the lumbar vertebrae. The sagittal projection is a two-dimensional vector composed of the forward / backward and up / down components of this vector. The angle between this two-dimensional vector and the vertically upward unit vector is calculated using inverse trigonometric functions. The left / right lateral tilt angle is calculated by the angle between the projection of the trunk axis vector onto the coronal plane and the vertical direction. The coronal projection is a two-dimensional vector composed of the left / right and up / down components of the trunk axis vector. The angle between this left / right lateral tilt angle and the vertically upward unit vector is calculated using inverse trigonometric functions.

[0100] The trunk tilt angle difference is obtained by subtracting the standard trunk tilt angle of the corresponding time frame in the standard movement template from the trunk tilt angle of each time frame in the current training movement. This difference also needs to be normalized to ensure that it is within the range of -180 degrees to +180 degrees. The normalization method is the same as the joint angle difference normalization method. The center of mass position coordinates are three-dimensional spatial coordinate values, representing the position of the center of mass in the forward, left-right, and up-down directions, respectively. The center of mass coordinates are calculated by weighted averaging of the mass of each limb segment of the astronaut's body. The position of the center of mass of each limb segment is estimated based on the joint coordinates and the limb segment mass distribution model.

[0101] The centroid position difference vector is obtained by subtracting the centroid position coordinates of each time frame in the current training motion from the standard centroid position coordinates of the corresponding time frame in the standard motion template according to each axis. This vector contains three components corresponding to the deviations in three spatial directions, with the unit being millimeters. The torso tilt angle difference contains two components: forward / backward and left / right. The centroid position difference vector contains three components: forward / backward, left / right, and up / down. These components are combined in each time frame to form a posture space deviation vector. This vector contains five components in order: the difference in torso forward / backward tilt angle, the difference in torso left / right tilt angle, the difference in centroid position in the forward / backward direction, the difference in centroid position in the left / right direction, and the difference in centroid position in the up / down direction. The first two components are in degrees, and the last three components are in millimeters.

[0102] The joint angle deviation matrix and attitude space deviation vector are weighted and fused according to preset dimensionality weighting coefficients to generate the final attitude deviation vector. The weighted fusion operation requires dimensionality reduction of the joint angle deviation matrix first. Dimensionality reduction is achieved by calculating the root mean square (RMS) values ​​for both the time and joint dimensions of the matrix. Specifically, the sum of squares of the three-degree-of-freedom angle differences for each joint across all time frames is calculated. This sum is accumulated by summing the squares of the pitch, yaw, and roll angle differences for that joint across all time frames. The sum is then divided by the number of time frames and multiplied by three to obtain the RMS value of the three-degree-of-freedom angle differences for that joint. Taking the square root of this RMS value yields the RMS angle deviation for that joint, reflecting the average deviation amplitude of the joint throughout the entire motion. The RMS angle deviations of all joints form the joint deviation vector, with a length equal to the number of joints, typically eighteen. The joint deviation vector is further weighted and averaged. The weighting coefficients are set according to the degree of influence of the joints on the quality of the movement. Core joints such as the shoulder, hip, and knee joints have a greater impact on the stability and accuracy of the movement, and their weighting coefficient is set to 0.3. Secondary joints such as the elbow and ankle joints have a weighting coefficient of 0.2. End joints such as the wrist and finger joints have a smaller impact on the overall movement, and their weighting coefficient is set to 0.1. The joint deviation vector is multiplied element-wise by its corresponding weighting coefficient and then summed. The summation result is divided by the sum of all weighting coefficients to obtain the weighted average joint angle deviation scalar value. This scalar value comprehensively reflects the overall deviation level of the joint angles throughout the body.

[0103] The five components of the posture space deviation vector also need to be multiplied by their corresponding dimension weighting coefficients to weight the importance of deviations in different dimensions. The difference in the torso's forward and backward tilt angles has the greatest impact on overall movement stability and center of gravity control, with a weighting coefficient of 0.4. The difference in the torso's left and right tilt angles has the next greatest impact on movement symmetry and balance, with a weighting coefficient of 0.3. The difference in the center of mass's forward and backward position reflects forward and backward movement deviation, with a weighting coefficient of 0.2. The differences in the center of mass's left and right position and up and down position reflect lateral movement deviation and vertical movement deviation, respectively, with weighting coefficients of 0.15 each. Multiplying each component by its corresponding weighting coefficient yields the weighted posture space deviation vector, which maintains a five-dimensional structure but whose component values ​​have been adjusted according to their importance. The weighted average joint angle deviation scalar value is used as the first component of the fused attitude deviation vector. The five components of the weighted attitude space deviation vector are used as the second to sixth components of the fused attitude deviation vector in turn. The final attitude deviation vector is a six-dimensional vector. The first component, in degrees, represents the comprehensive deviation of joint angles. The second and third components, in degrees, represent the deviation of torso tilt angles. The fourth, fifth, and sixth components, in millimeters, represent the deviation of the center of mass position.

[0104] Taking astronauts performing squat training as an example, the implementation process is illustrated. The current training sequence contains 120 frames of data with a sampling frequency of 60 frames per second, corresponding to a 2-second action duration. The standard squat template contains 100 frames of data with the same sampling frequency of 60 frames per second. The dynamic time warping algorithm constructs a cost matrix of 120 rows and 100 columns. Taking the right knee joint as an example, in the 40th frame of the current action, the right knee pitch angle is -95.3 degrees, the yaw angle is 2.1 degrees, and the roll angle is -1.5 degrees. In the corresponding mapped frame of the standard template, the right knee pitch angle is -90 degrees, the yaw angle is 0 degrees, and the roll angle is 0 degrees. The sum of the squares of the three degrees of freedom angle differences of the right knee joint in this frame is 5.3 squared plus 2.1 squared plus 1.5 squared, which equals 33.4. The square root of the sum of the squares of the differences of all 18 joints is used to obtain the inter-frame distance of this frame. After the cost matrix is ​​filled, the optimal path length is found to be 130 steps. The path maps the current action from frames 1 to 10 to frame 1 of the standard template, frames 11 to 80 to frames 2 to 70 of the standard template, and frames 81 to 120 to frames 71 to 100 of the standard template. Based on the mapping relationship, the joint angle deviation matrix is ​​calculated. The sum of squares of the three-degree-of-freedom angle differences of the right knee joint in 120 frames is 3205. Dividing by 360 gives a mean square value of 8.9. Taking the square root gives a root mean square angle deviation of 2.98 degrees. The root mean square angle deviations of the eighteen joints are calculated separately and multiplied by the corresponding weight coefficients. The root mean square deviations of the core joints, such as the right knee, left knee, right hip, and left hip, are 2.98 degrees, 2.85 degrees, 2.3 degrees, and 2.1 degrees, respectively, with a weight of 0.3. The root mean square deviations of the elbow and ankle joints are approximately 1.8 degrees with a weight of 0.2, and the root mean square deviations of the wrist and fingers are approximately 0.9 degrees with a weight of 0.1. After weighted summation, the weighted average joint angle deviation is obtained by dividing by the total weight, which is 2.15 degrees. The root mean square (RMS) values ​​for the torso's forward / backward tilt angle (RMS) at 120 frames are 1.9 degrees multiplied by a weight of 0.4, resulting in 0.76 degrees. The RMS value for the left / right tilt angle (RMS) is 0.8 degrees multiplied by a weight of 0.3, resulting in 0.24 degrees. The RMS value for the center of mass's forward / backward position (RMS) is 8 millimeters multiplied by a weight of 0.2, resulting in 1.6 millimeters. The RMS value for the left / right position (RMS) is 12 millimeters multiplied by a weight of 0.15, resulting in 1.8 millimeters. The RMS value for the vertical position (RMS) is 20 millimeters multiplied by a weight of 0.15, resulting in 3 millimeters. The final posture deviation vector is 2.15 degrees, 0.76 degrees, 0.24 degrees, 1.6 millimeters, 1.8 millimeters, and 3 millimeters. This vector comprehensively reflects the multi-dimensional deviation between the current squatting motion and the standard motion.

[0105] In one optional implementation, based on the posture deviation vector, the target joint angle or target limb spatial position that the deviation dimension needs to achieve is calculated, the optimal adjustment path from the current posture state to the target joint angle or the target limb spatial position is planned, and posture correction instructions including step-by-step execution instructions and execution timing control are generated, including:

[0106] Based on the joint angle deviation value or limb spatial position deviation value corresponding to the deviation dimension in the posture deviation vector, and combined with the standard joint angle sequence and standard limb spatial posture parameters in the standard action template for aerospace missions, the target joint angle or target limb spatial position that the deviation dimension needs to achieve is calculated, and a graded correction target set is generated.

[0107] Based on each correction target in the graded correction target set, the tolerable adjustment rate of the joint or limb is determined according to the heart rate index in the astronaut's current physiological state parameters, and the optimal adjustment path from the current posture state to the target joint angle or the target limb spatial position is planned;

[0108] The intermediate posture node sequence in the optimal adjustment path is converted into specific joint rotation direction indicators, joint rotation amplitude values, and limb movement direction vectors, generating posture correction instructions that include step-by-step execution instructions and execution timing control.

[0109] When calculating the target joint angle or target limb spatial position based on the attitude deviation vector, the specific deviation values ​​for each deviation dimension need to be extracted from the attitude deviation vector. The attitude deviation vector contains six components: the first component is a weighted average joint angle deviation scalar value in degrees; the second and third components are the trunk forward / backward tilt angle deviation and left / right lateral tilt angle deviation, respectively, in degrees; and the fourth, fifth, and sixth components are the center of mass position deviations in the forward / backward, left / right, up / down directions, respectively, in millimeters. For the joint angle deviation dimension, the three-degree-of-freedom angle deviation value of each joint in the current time frame is extracted from the joint angle deviation matrix. This deviation value represents the difference between the joint angle in the current training action and the corresponding joint angle in the standard action template. Combining the standard joint angle sequence in the aerospace mission standard action template, the standard joint angle in the corresponding time frame of the standard joint angle sequence is taken as the target joint angle, or the current joint angle is subtracted from the angle deviation value to obtain the target joint angle; the two calculation methods are equivalent. For the limb spatial position deviation dimension, the trunk tilt angle deviation and center of mass position deviation are extracted. Combining the standard trunk tilt angle and standard center of mass position coordinates in the standard limb spatial posture parameters, the standard values ​​are taken as the target limb spatial position.

[0110] When generating a tiered set of correction targets, the targets are prioritized based on the magnitude and importance of their deviation values. Deviation values ​​are measured in absolute value: joint angle deviations greater than 10 degrees are defined as high-priority targets, deviations between 5 and 10 degrees as medium-priority targets, and deviations less than 5 degrees as low-priority targets. Similarly, trunk tilt angle deviations greater than 8 degrees are defined as high-priority, deviations between 4 and 8 degrees as medium-priority, and deviations less than 4 degrees as low-priority. Center of mass position deviations greater than 50 mm in the forward / backward or left / right directions are defined as high-priority, deviations between 20 and 50 mm as medium-priority, and deviations less than 20 mm as low-priority. The importance of deviations is reflected by dimensional weighting coefficients: the forward / backward tilt angle deviation has the highest priority at 0.4, followed by the left / right tilt angle deviation at 0.3, the forward / backward center of mass deviation at 0.2, and the left / right and up / down center of mass deviations at 0.15. The importance of joint angle deviation is distinguished by the joint weight coefficient. Core joints such as the hip, knee and shoulder joints have a weight of 0.3 and a priority higher than secondary joints such as the elbow and ankle joints with a weight of 0.2. Terminal joints such as the wrist joint have a weight of 0.1 and the lowest priority.

[0111] By combining the magnitude of the deviation value with its importance weight, all correction targets are sorted from largest to smallest according to their weighted deviation values ​​(calculated by multiplying the absolute value of the deviation value by a weighting coefficient), forming a hierarchical correction target set. Each element in the hierarchical correction target set includes the correction target type (e.g., joint angle or limb position), target joint or limb name, target angle value or target position coordinates, current angle value or current position coordinates, deviation value, priority level, and weighted deviation value. The set is arranged in descending order of weighted deviation value to ensure that the correction targets with the largest and most important deviations are processed first. When planning the optimal adjustment path from the current attitude state to the target joint angle or target limb spatial position based on each correction target in the hierarchical correction target set, the tolerable adjustment rate of the joint or limb needs to be determined based on the astronaut's current physiological parameters, specifically the heart rate. Heart rate reflects the astronaut's current exercise intensity and fatigue level. A heart rate below 100 beats per minute indicates that the astronaut is in a low-intensity exercise state and can withstand a relatively fast adjustment rate. A heart rate between 100 and 140 beats per minute indicates a moderate-intensity exercise state and can withstand a moderate adjustment rate. A heart rate above 140 beats per minute indicates a high-intensity exercise state or a fatigue state, requiring a reduction in the adjustment rate to avoid excessive load.

[0112] The tolerable adjustment rate is defined as the maximum change in joint angle or limb position per unit time. The unit for joint angle adjustment rate is degrees per second, and the unit for limb position adjustment rate is millimeters per second. When the heart rate is below 100 beats per minute, the upper limit for joint angle adjustment rate is set at 30 degrees per second, and the upper limit for limb position adjustment rate is set at 100 millimeters per second. When the heart rate is between 100 and 140 beats per minute, the upper limit for joint angle adjustment rate is set at 20 degrees per second, and the upper limit for limb position adjustment rate is set at 70 millimeters per second. When the heart rate is above 140 beats per minute, the upper limit for joint angle adjustment rate is set at 10 degrees per second, and the upper limit for limb position adjustment rate is set at 40 millimeters per second. Optimal adjustment path planning is implemented using the A* algorithm, which searches the state space for the shortest path from the current posture state to the target posture state. The state space is defined as a high-dimensional space consisting of all possible combinations of joint angles or limb positions, where each state node represents a specific posture configuration.

[0113] The A* algorithm maintains two data structures: an open list and a closed list. The open list stores state nodes to be explored, using a min-heap data structure for efficient extraction of the minimum-cost node. The closed list stores explored state nodes, using a hash table for fast lookup and to avoid redundant expansion. At startup, the current pose state is added to the open list as the starting node. The actual cost of the starting node is set to zero, and the estimated cost is set to the heuristic distance from the starting node to the target pose state. The heuristic distance is calculated as the sum of the absolute differences between all joint angles in the current state and the target joint angles, plus the Euclidean distance between all limb position coordinates and the target position coordinates. This heuristic function satisfies the admissibility condition, meaning it does not overestimate the actual shortest path cost, ensuring that the path found by the A* algorithm is the optimal path. The algorithm selects the node with the minimum total cost from the open list for expansion; the total cost equals the actual cost plus the estimated cost.

[0114] When a node is expanded, all its neighboring nodes are generated. A neighboring node is defined as a new state node obtained by changing a joint angle or limb position coordinate along a certain direction by a fixed step size based on the current node. The step size for changing joint angles is set to five degrees, and the step size for changing limb position coordinates is set to twenty millimeters. The direction of change includes both increasing and decreasing. When generating neighboring nodes, it is necessary to check whether the node satisfies joint angle constraints and kinematic constraints. Joint angle constraints require that the rotation angle of each joint be within the physiological range of the joint. For example, the range of the knee joint's pitch angle is -150 degrees to 0 degrees, and the range of the elbow joint's pitch angle is 0 degrees to 150 degrees. Nodes outside these ranges are not added to the expansion set. Kinematic constraints require that the limb segment length remain constant during the adjustment process; that is, the Euclidean distance between adjacent joints is equal to the bone length of that limb segment. Nodes with a deviation exceeding five percent of the bone length are considered to violate the constraints and are not added to the expansion set.

[0115] For each neighboring node satisfying the constraints, calculate the actual cost of reaching that neighboring node from the starting node through the current node. The actual cost equals the actual cost of the current node plus the movement cost from the current node to the neighboring node. The movement cost calculation considers the adjustment rate constraint. If the joint angle change step size is five degrees and the adjustment rate limit is thirty degrees per second, then the movement cost is five divided by thirty, which equals 0.167 seconds. The movement cost is further multiplied by a weight coefficient reflecting the importance of the joint or limb: the weight of a core joint is 0.5, the weight of a secondary joint is 0.5, and the weight of an distal joint is 0.3. Calculate the estimated cost of the neighboring node as the heuristic distance from that node to the target posture state. Add the actual cost to the estimated cost to obtain the total cost. Check if the neighboring node already exists in the open or closed list. If the neighboring node is not in either list, add it to the open list and record its parent node as the current node. If the neighboring node is already in the open list and the newly calculated actual cost is less than the original actual cost, update the actual cost, total cost, and parent node of that node.

[0116] If an adjacent node is already in the closed list but the newly calculated actual cost value is less than the original actual cost value, the node needs to be removed from the closed list and re-added to the open list, and its cost value and parent node updated. After the current node is expanded, it is removed from the open list and added to the closed list. The next round of expansion continues by selecting the node with the smallest total cost value from the open list. The algorithm terminates when the expanded node is the target attitude state node. A complete path is constructed from the target node to the starting node by backtracking the parent node chain. This path is the optimal adjustment path, represented as an intermediate attitude node sequence. When converting the intermediate attitude node sequence in the optimal adjustment path into specific joint rotation direction indicators, joint rotation amplitude values, and limb movement direction vectors, the state changes between adjacent nodes in the node sequence are calculated. For joint angle adjustment, the three degrees of freedom rotation angle values ​​of the same joint in two adjacent nodes are compared. If the angle value of a certain degree of freedom increases, the rotation direction indicator is positive rotation; if the angle value decreases, the rotation direction indicator is negative rotation; if the angle value remains unchanged, there is no need to adjust that degree of freedom.

[0117] The joint rotation amplitude is equal to the absolute difference in angle values ​​of that degree of freedom between adjacent nodes, in degrees. For limb position adjustment, the coordinates of the center of mass or the torso tilt angle of the same limb in two adjacent nodes are compared, and a coordinate difference vector is calculated. The three components of the vector represent the positional changes in the forward, backward, left, right, up, and down directions, respectively. This vector is the limb movement direction vector. When generating step-by-step execution instructions, the adjustment operation between each pair of adjacent nodes is encoded into one instruction. The instruction includes the adjustment object identifier, adjustment type, adjustment direction, adjustment amplitude, and execution timing control parameters. The execution timing control parameters are calculated based on the upper limit of the adjustment rate and the adjustment amplitude. The instruction duration is equal to the adjustment amplitude divided by the upper limit of the adjustment rate. The instruction execution time is calculated by accumulating the duration of the preceding instructions. The execution time of the first instruction is zero, and the execution time of subsequent instructions is the sum of the durations of the preceding instructions.

[0118] If multiple joints or limbs can be adjusted in parallel, the execution time of the corresponding instructions can be set to the same value to achieve synchronous execution. The implementation process is illustrated using the correction of the right knee joint angle deviation during a squatting motion by an astronaut as an example. The right knee joint pitch angle deviation extracted from the attitude deviation vector is -5.3 degrees, the current right knee joint pitch angle is -95.3 degrees, the standard right knee joint pitch angle at the corresponding moment in the standard action template is -90 degrees, and the target joint angle is calculated to be -90 degrees. The astronaut's current heart rate is 120 beats per minute, indicating moderate-intensity exercise, and the upper limit of the joint angle adjustment rate is set to 20 degrees per second. When the A* algorithm plans the adjustment path, the starting node state is the current right knee pitch angle of -95.3 degrees, the target node state is the target pitch angle of -90 degrees, and the heuristic distance is 5.3 degrees. The algorithm expands the starting node to generate two adjacent nodes. The heuristic distance between nodes with an increased pitch angle of 5 degrees to -90.3 degrees is 0.3 degrees, the actual cost is 0.25 seconds, and the total cost is 0.55. The heuristic distance between nodes with an increased pitch angle of 5 degrees to -100.3 degrees is 10.3 degrees, and the total cost is 10.55.

[0119] The algorithm selects the node with the lowest total value, i.e., the node with a pitch angle of -90.3 degrees, and continues to expand. A variable step size mechanism is used during this expansion; when the distance between the node and the target is less than the fixed step size, the step size is adjusted to the actual distance value of 0.3 degrees, generating a target node with a pitch angle of -90 degrees. The algorithm terminates the search upon reaching the target node, and the backtracking path yields the intermediate attitude node sequence: pitch angles of -95.3 degrees, -90.3 degrees, and -90 degrees. When converted into step-by-step execution instructions, the first step adjusts the right knee joint, with the adjustment type being angle rotation, the adjustment direction being positive pitch rotation, and the adjustment amplitude being 5 degrees. The execution time is 0 seconds, and the duration is 0.25 seconds. The second step adjusts the right knee joint again, with the adjustment type being angle rotation, the adjustment direction being positive pitch rotation, and the adjustment amplitude being 0.3 degrees. The execution time is 0.25 seconds, and the duration is 0.015 seconds, forming the right knee joint attitude correction instruction sequence. The total execution time is 0.265 seconds.

[0120] In one optional implementation, the tolerable adjustment rate of the joint or limb is determined based on the heart rate index in the astronaut's current physiological state parameters, and the optimal adjustment path from the current posture state to the target joint angle or the target limb spatial position is planned, including:

[0121] Based on heart rate indicators, the cardiovascular load state is determined to belong to the load zone, and a corresponding rate adjustment coefficient is assigned to the adjustment movement of the joint or limb. The rate adjustment coefficient is used to regulate the execution speed of the joint or limb adjusting from the current posture state to the target joint angle or the target limb spatial position.

[0122] Calculate the total angular change or total spatial displacement of the joint or limb from its current posture to the target joint angle or the target limb's spatial position. Divide the total angular change or total spatial displacement into segments according to the rate adjustment coefficient to generate multiple sub-adjustment angles or sub-adjustment displacements.

[0123] The endpoint of each sub-adjustment angle or sub-adjustment displacement is taken as an intermediate posture node. The intermediate posture nodes are sequentially set on the adjustment trajectory from the current posture state to the target joint angle or the target limb spatial position to form an intermediate posture node sequence.

[0124] For the sub-adjustment angle or sub-adjustment displacement between two adjacent intermediate attitude nodes in the intermediate attitude node sequence, the execution time between nodes required to complete the sub-adjustment angle or sub-adjustment displacement is calculated in conjunction with the rate adjustment coefficient;

[0125] The intermediate posture node sequence is combined with the execution time between each node to generate the optimal adjustment path from the current posture state to the target joint angle or the target limb spatial position.

[0126] The determination of cardiovascular load status is based on real-time monitoring of heart rate indicators and a zone classification mechanism. Astronaut heart rate data is collected at a frequency of 100 Hz using a chest heart rate sensor, and a 5-point moving average filter is used to eliminate sampling noise. Heart rate load status is divided into five levels: low load zone (heart rate less than resting heart rate + 15 beats per minute), mild load zone (heart rate between resting heart rate + 15 and 30 beats per minute), moderate load zone (heart rate between resting heart rate + 30 and 45 beats per minute), heavy load zone (heart rate between resting heart rate + 45 and 60 beats per minute), and extremely heavy load zone (heart rate exceeding resting heart rate + 60 beats per minute). Resting heart rate data is obtained from the astronaut's personal physiological records. The system supports periodic calibration and updates of resting heart rate every 7 days, with an accuracy requirement of 1 time per minute. Anomaly detection of heart rate data is achieved through upper and lower bound limits. Anomaly alarms and data re-acquisition procedures are triggered when the heart rate is below 40 beats per minute or above 200 beats per minute.

[0127] The rate adjustment coefficient is allocated based on the safety assessment principles of cardiovascular load status. A rate adjustment coefficient of 1.0 is assigned to the low-load zone, indicating that adjustments can be performed at the standard rate. The rate adjustment coefficient is assigned 0.8 to the light-load zone, 0.6 to the moderate-load zone, 0.4 to the heavy-load zone, and 0.2 to the extremely heavy-load zone. The rate adjustment coefficient is a floating-point number with a precision of 0.05. The minimum value is limited to 0.1 to avoid excessively slow adjustment, and the maximum value is limited to 1.0 to prevent over-adjustment. The system maintains a joint importance weight table. The rate adjustment coefficient for safety-critical joints such as the spine, hip, and knee is multiplied by an additional safety factor of 0.9; for motion-critical joints such as the shoulder and elbow, it is multiplied by a factor of 0.95; and for secondary joints such as the fingers and toes, the original coefficient remains unchanged. The rate adjustment coefficient is updated when the heart rate changes by more than 5 beats per minute or when the load status changes across zones.

[0128] The total angular change is calculated by the vector difference between the current joint angle and the target joint angle. The joint angle uses the ZYX Euler angle representation, which includes the yaw angle around the Z-axis, the pitch angle around the Y-axis, and the roll angle around the X-axis. The system calculates the angle difference along the three axes separately, takes the absolute value, and selects the maximum value as the total angular change of the joint. The calculation of the angle difference takes into account the periodicity of the angle, ensuring that the difference is within the range of -180 degrees to +180 degrees. The total spatial displacement is calculated based on the three-dimensional vector distance of the limb's center of mass coordinates, using the Euclidean distance formula, which is the square root of the sum of the squares of the X-axis displacement, the squares of the Y-axis displacement, and the squares of the Z-axis displacement. The limb coordinate system has the astronaut's torso center of mass as the origin, with the positive X-axis pointing forward, the positive Y-axis pointing to the right, and the positive Z-axis pointing upward. The coordinate accuracy is at the millimeter level, and the measurement error is controlled within 2 millimeters.

[0129] The segmented division is based on the reciprocal of the rate adjustment coefficient to determine the number of sub-adjustments. The number of sub-adjustments equals the integer part of the reciprocal of the rate adjustment coefficient plus 1. When the rate adjustment coefficient is 1.0, the number of sub-adjustments is 2, and the total change is divided into 2 segments. When the rate adjustment coefficient is 0.8, the number of sub-adjustments is 2; when the rate adjustment coefficient is 0.6, it is 2; when the rate adjustment coefficient is 0.4, it is 3; and when the rate adjustment coefficient is 0.2, it is 6. The formula for calculating the number of segments is to round up the result of 1 divided by the rate adjustment coefficient. The maximum number of segments is limited to 20, and the minimum number of segments is limited to 2. Each sub-adjustment angle is calculated by dividing the total angle change by the number of segments. The minimum value of the sub-adjustment angle is limited to 1 degree, and the minimum value of the sub-adjustment displacement is limited to 5 millimeters to avoid system oscillations caused by fine-tuning.

[0130] Intermediate attitude nodes are generated based on the cumulative sequence of sub-adjustment angles or sub-adjustment displacements. Starting from the current attitude state, each attitude state after completing a sub-adjustment is considered an intermediate attitude node. The data structure of an intermediate attitude node includes four fields: node number, joint angle array, limb position array, and node timestamp. The joint angle array stores the angle information of 23 major joints, each joint containing angle values ​​for 3 degrees of freedom, in 32-bit floating-point format. The limb position array stores the 3D coordinate positions of six limbs: head, torso, left arm, right arm, left leg, and right leg, in 32-bit floating-point format. The node number uses an incrementing integer sequence, and the node timestamp records the system time of node generation with millisecond precision. The intermediate attitude node sequence is stored using a dynamic array, supporting random access and sequential traversal operations.

[0131] The calculation of execution time between nodes is based on a comprehensive calculation of the sub-adjustment amount, the standard adjustment speed, and the rate adjustment coefficient. Execution time equals the sub-adjustment amount divided by the standard adjustment speed, then divided by the rate adjustment coefficient. The standard adjustment speed for joint rotation is set to 30 degrees per second, and the standard adjustment speed for limb movement is set to 40 millimeters per second. The data precision for execution time is 0.01 seconds, with a minimum execution time limit of 0.05 seconds to avoid instantaneous adjustments and a maximum execution time limit of 3 seconds to prevent excessively slow single-step adjustments. The system supports variable speed adjustment modes, including four options: constant speed mode, linear acceleration mode, linear deceleration mode, and sine curve mode. In constant speed mode, the execution time between nodes remains constant; in linear acceleration mode, the execution time gradually decreases; in linear deceleration mode, the execution time gradually increases; and in sine curve mode, the execution time changes according to a sine function to achieve a smooth acceleration and deceleration effect.

[0132] The optimal adjustment path is generated through data binding between intermediate attitude node sequences and corresponding execution duration sequences. The path data structure includes seven core fields: path identifier, starting node, target node, intermediate node array, duration array, total execution time, and path status. The path identifier uses a 128-bit UUID format to ensure global uniqueness. The starting node and target node store complete information about the current and target attitude states, respectively. The intermediate node array stores all intermediate attitude nodes in chronological order, the duration array stores the execution duration data between adjacent nodes, and the total execution time is the sum of the durations between all nodes. The path status field identifies the current execution status of the path, including five status values: not started, in progress, completed, paused, and terminated. The system supports serialized storage and network transmission of path data, using JSON format to encapsulate path information with a compression rate exceeding 40%.

[0133] Path verification and optimization comprises three stages: continuity check, boundary constraint check, and smoothness optimization. The continuity check verifies that the angle and displacement changes between adjacent nodes are within reasonable ranges, with a threshold of 45 degrees per second for angle change rate and 60 millimeters per second for displacement change rate. The boundary constraint check ensures that the joint angles of all intermediate attitude nodes are within physiological ranges and that limb positions are within reachable space within the spacecraft. Smoothness optimization uses a cubic spline interpolation algorithm to curve-fit the path, eliminating sharp inflections and discontinuous jumps, achieving an interpolation accuracy five times that of the original sampling point density.

[0134] This practical application case demonstrates the adjustment process of the right shoulder joint's pitch angle. The current angle is 15 degrees, the target angle is 75 degrees, and the total angle change is 60 degrees. The heart rate sensor detects a current heart rate of 78 beats per minute and a resting heart rate of 62 beats per minute. The heart rate exceeds the resting heart rate by 16 beats per minute, which falls within the light load zone, and the rate adjustment coefficient is 0.8. Following the segmentation rule, the 60-degree total angle change is divided into two sub-adjustment angles, each 30 degrees. The intermediate posture node sequence contains three nodes with angles of 15 degrees, 45 degrees, and 75 degrees respectively. The standard joint rotation speed is 30 degrees per second; combined with the rate adjustment coefficient of 0.8, the actual adjustment speed is 24 degrees per second. The execution time for the first 30-degree sub-adjustment is 1.25 seconds, and the execution time for the second 30-degree sub-adjustment is also 1.25 seconds, for a total execution time of 2.5 seconds. The generated optimal adjustment path contains a complete trajectory sequence of the initial state, intermediate nodes, and target state. The total number of nodes in the path is 3, and the total execution time is 2.5 seconds. The path verification passes the continuity and boundary constraint checks, and the smoothness-optimized path curve conforms to physiological motion characteristics.

[0135] A second aspect of the present invention provides an intelligent guidance and feedback system for astronaut physical training, comprising:

[0136] The first unit is used to acquire real-time three-dimensional motion trajectory data of astronauts at multiple joints during training through a space-distributed motion capture device.

[0137] The second unit is used to perform spatial clustering on the multi-joint three-dimensional motion trajectory data according to human anatomy partitioning rules to obtain multiple joint set, construct multiple kinematic skeletal chains according to hierarchical parent-child relationships, and perform inverse kinematic analysis calculations sequentially from the terminal joints along the skeletal chains to the root joints to obtain the joint angle sequence and limb spatial posture parameters of the astronaut's current training movements;

[0138] The third unit is used to perform multi-dimensional spatial alignment between the astronaut's current training movements and the preset space mission standard movement template based on the joint angle sequence and the limb spatial posture parameters, so as to obtain the posture deviation vector;

[0139] The fourth unit is used to calculate the target joint angle or target limb spatial position that the deviation dimension needs to reach based on the posture deviation vector, plan the optimal adjustment path from the current posture state to the target joint angle or target limb spatial position, and generate posture correction instructions that include step-by-step execution instructions and execution timing control;

[0140] The fifth unit is used to output attitude correction commands to astronauts through a real-time feedback device and to simultaneously adjust the drag load parameters of the training equipment.

[0141] A third aspect of the present invention provides an electronic device, comprising:

[0142] processor;

[0143] Memory used to store processor-executable instructions;

[0144] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0145] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0146] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent guidance and feedback method for astronaut physical training, characterized in that, include: Real-time three-dimensional motion trajectory data of astronauts at multiple joint points during training is acquired using a space-distributed motion capture device. The multi-joint 3D motion trajectory data is spatially clustered according to human anatomical zoning rules to obtain multiple joint set sets. Multiple kinematic skeletal chains are constructed according to hierarchical parent-child relationships. Inverse kinematic analysis is performed sequentially from the terminal joints along the skeletal chains towards the root joints to obtain the joint angle sequence and limb spatial posture parameters of the astronaut's current training movement, including: The multi-joint 3D motion trajectory data is spatially clustered according to human anatomical zoning rules to obtain multiple joint set sets; Based on predefined skeletal connection relationships, multiple kinematic skeleton chains are constructed from the joints in the multiple sets of joints according to hierarchical parent-child relationships. Each kinematic skeleton chain contains a root joint and an end joint. For each kinematic skeletal chain, inverse kinematic analysis is performed sequentially from the terminal joint along the skeletal chain towards the root joint. Based on the three-dimensional spatial positional relationship between adjacent joints and the bone length constraint, the three-degree-of-freedom rotation angles of each joint are solved to obtain a joint angle sequence; where... Starting from the terminal joint, each joint is traversed sequentially along the kinematic skeleton chain toward the root joint. For the currently traversed joint and its parent joint, a connection vector is calculated based on the three-dimensional spatial positional relationship between them. The connection vector is then normalized in combination with the bone length constraint to obtain a unit direction vector sequence. For each unit direction vector in the unit direction vector sequence, the unit direction vector is decomposed into three orthogonal rotation components relative to the local coordinate system of the parent joint. By calculating the angles between the three orthogonal rotation components and each coordinate axis of the local coordinate system of the parent joint, the rotation angles of each joint in the pitch, yaw, and roll directions are solved to obtain the joint angle sequence. The three-degree-of-freedom rotation angle of each joint in the joint angle sequence is compared one by one with the preset joint physiological range of motion constraints. When a rotation angle that exceeds the joint physiological range of motion constraints is detected, an angle correction mechanism is triggered. For rotation angles that exceed the physiological range of motion of the joint, the maximum allowable rotation angle of the joint is calculated based on the spacesuit's limiting constraints, and the excess rotation angle is redistributed to the joints adjacent to the joint in the kinematic skeletal chain, thus updating the joint angle sequence. Based on the angle values ​​of the root joints in the joint angle sequence and the overall spatial orientation of the kinematic skeletal chain, the tilt angle of the astronaut's torso relative to the direction of gravity and the coordinates of the center of mass are calculated to obtain the spatial posture parameters of the limbs. Based on the joint angle sequence and the limb spatial posture parameters, the astronaut's current training movements are aligned with the preset space mission standard movement template in a multi-dimensional space to obtain the posture deviation vector; Based on the posture deviation vector, calculate the target joint angle or target limb spatial position that the deviation dimension needs to achieve, plan the optimal adjustment path from the current posture state to the target joint angle or target limb spatial position, and generate posture correction instructions that include step-by-step execution instructions and execution timing control; The system outputs attitude correction commands to the astronauts through a real-time feedback device and simultaneously adjusts the drag load parameters of the training equipment.

2. The method according to claim 1, characterized in that, Based on the joint angle sequence and the limb spatial posture parameters, the astronaut's current training movements are aligned with the preset space mission standard movement template in a multi-dimensional space, resulting in a posture deviation vector including: Align the joint angle sequence with the standard joint angle sequence in the standard action template for aerospace missions along the time axis to establish a temporal mapping relationship between the current training action and the standard action template for aerospace missions. Based on the aforementioned temporal mapping relationship, for each joint in the joint angle sequence, the angle difference between its three-degree-of-freedom rotation angle and the corresponding rotation angle of the corresponding joint at the corresponding time in the standard joint angle sequence is calculated to obtain the joint angle deviation matrix; For the trunk tilt angle and the center of mass position coordinates in the limb spatial posture parameters, calculate their spatial differences with the corresponding parameters in the standard limb spatial posture parameters to obtain the posture spatial deviation vector; The joint angle deviation matrix and the attitude space deviation vector are weighted and fused according to a preset dimension weight coefficient to generate the attitude deviation vector.

3. The method according to claim 1, characterized in that, Based on the posture deviation vector, calculate the target joint angle or target limb spatial position that the deviation dimension needs to achieve, plan the optimal adjustment path from the current posture state to the target joint angle or target limb spatial position, and generate posture correction instructions that include step-by-step execution instructions and execution timing control, including: Based on the joint angle deviation value or limb spatial position deviation value corresponding to the deviation dimension in the posture deviation vector, and combined with the standard joint angle sequence and standard limb spatial posture parameters in the standard action template for aerospace missions, the target joint angle or target limb spatial position that the deviation dimension needs to achieve is calculated, and a graded correction target set is generated. Based on each correction target in the graded correction target set, the tolerable adjustment rate of the joint or limb is determined according to the heart rate index in the astronaut's current physiological state parameters, and the optimal adjustment path from the current posture state to the target joint angle or the target limb spatial position is planned; The intermediate posture node sequence in the optimal adjustment path is converted into specific joint rotation direction indicators, joint rotation amplitude values, and limb movement direction vectors, generating posture correction instructions that include step-by-step execution instructions and execution timing control.

4. The method according to claim 3, characterized in that, Based on the heart rate index in the astronaut's current physiological parameters, the tolerable adjustment rate of the joint or limb is determined, and the optimal adjustment path from the current posture to the target joint angle or the target limb spatial position is planned, including: Based on heart rate indicators, the cardiovascular load state is determined to belong to the load zone, and a corresponding rate adjustment coefficient is assigned to the adjustment movement of the joint or limb. The rate adjustment coefficient is used to regulate the execution speed of the joint or limb adjusting from the current posture state to the target joint angle or the target limb spatial position. Calculate the total angular change or total spatial displacement of the joint or limb from its current posture to the target joint angle or the target limb's spatial position. Divide the total angular change or total spatial displacement into segments according to the rate adjustment coefficient to generate multiple sub-adjustment angles or sub-adjustment displacements. The endpoint of each sub-adjustment angle or sub-adjustment displacement is taken as an intermediate posture node. The intermediate posture nodes are sequentially set on the adjustment trajectory from the current posture state to the target joint angle or the target limb spatial position to form an intermediate posture node sequence. For the sub-adjustment angle or sub-adjustment displacement between two adjacent intermediate attitude nodes in the intermediate attitude node sequence, the execution time between nodes required to complete the sub-adjustment angle or sub-adjustment displacement is calculated in conjunction with the rate adjustment coefficient; The intermediate posture node sequence is combined with the execution time between each node to generate the optimal adjustment path from the current posture state to the target joint angle or the target limb spatial position.

5. An intelligent guidance and feedback system for astronaut physical training, used to implement the method as described in any one of claims 1-4, characterized in that, include: The first unit is used to acquire real-time three-dimensional motion trajectory data of astronauts at multiple joints during training through a space-distributed motion capture device. The second unit is used to perform spatial clustering on the multi-joint three-dimensional motion trajectory data according to human anatomy partitioning rules to obtain multiple joint set, construct multiple kinematic skeletal chains according to hierarchical parent-child relationships, and perform inverse kinematic analysis calculations sequentially from the terminal joints along the skeletal chains to the root joints to obtain the joint angle sequence and limb spatial posture parameters of the astronaut's current training movements; The third unit is used to perform multi-dimensional spatial alignment between the astronaut's current training movements and the preset space mission standard movement template based on the joint angle sequence and the limb spatial posture parameters, so as to obtain the posture deviation vector; The fourth unit is used to calculate the target joint angle or target limb spatial position that the deviation dimension needs to reach based on the posture deviation vector, plan the optimal adjustment path from the current posture state to the target joint angle or target limb spatial position, and generate posture correction instructions that include step-by-step execution instructions and execution timing control; The fifth unit is used to output attitude correction commands to astronauts through a real-time feedback device and to simultaneously adjust the drag load parameters of the training equipment.

6. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 4.

7. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 4.

Images