Joint track dynamic compensation intelligent analysis device for competitive sports

Abstract

The invention relates to the technical field of joint compensation, in particular to a competitive sports-oriented joint trajectory dynamic compensation intelligent analysis device, which comprises a torque fluctuation identification module, a path sudden change switching module, an impact interference identification module, a traction control adjustment module and a trajectory compensation control integration module. According to the method, take-off action stages are divided through joint torque state changes, a trajectory compensation structure is constructed by combining direction offset points in a supporting path, impact interference fragments are extracted based on posture sudden change features, and an action response process is adjusted according to the sequence difference of starting and ending positions in a traction path. The connection process of the trajectory segments is propelled by means of the arrangement relation of the execution nodes, a continuous path system covering a plurality of action conversion nodes is constructed, and the structural matching capacity for direction sudden change, impact disturbance and response delay in the action process is enhanced. And the path ductility, the action coordination and the connection stability of the joint track under the unsteady state motion condition are improved.

Description

Technical Field

[0001] This invention relates to the field of joint compensation technology, and in particular to an intelligent analysis device for dynamic compensation of joint trajectory in competitive sports. Background Technology

[0002] The field of joint compensation technology involves technologies for monitoring, analyzing, and regulating joint movement trajectories, angle changes, stress responses, and muscle synergy mechanisms during human movement. It primarily includes core aspects such as joint range of motion modeling, biomechanical modeling, motion capture analysis, sensor data acquisition and fusion, abnormal trajectory identification, and compensation strategy construction. This technology is widely applied in rehabilitation medicine, sports training, intelligent prosthetic control, and ergonomic equipment development, forming a systematic technical system integrating sensing, data processing, mechanical analysis, and feedback control. Among these, the traditional intelligent analysis device for dynamic compensation of joint trajectory in competitive sports refers to a device used to collect joint movement trajectory data during training or competition of high-level athletes, and to model, analyze, and compensate for trajectory deviations. This patent primarily addresses the nonlinear motion characteristics and trajectory deviation problems of human joints under high-intensity, rapidly changing conditions during movement. Traditional technologies typically involve deploying inertial measurement units and calculating joint displacement trajectories based on multi-axis angular velocity and acceleration data, combined with a feedforward motion model to construct correction curves to achieve approximate compensation and analysis of joint movement.

[0003] Existing technologies rely on trajectory calculation based on inertial measurement units during motion recognition, which suffers from problems such as trajectory response lag and insufficient state recognition accuracy. In complex motion states, it is difficult to distinguish the torque state changes of key nodes in the motion, and it is easy to ignore the path deviation links in the continuity of the motion. As a result, the trajectory compensation results lack accurate definition of directional changes and state disturbances. The connection logic of each stage within the motion sequence is unclear, which can easily lead to path breakpoints or control delays. In scenarios with drastic changes in motion trajectory, it is difficult to guarantee the response continuity and execution stability of the joint control structure, which affects the efficiency of identifying abnormal behavior in dynamic motion states and the structural reliability of compensation measures. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide an intelligent analysis device for dynamic compensation of joint trajectory in competitive sports.

[0005] On the one hand, an intelligent analysis device for joint trajectory dynamic compensation in competitive sports is provided, which includes: The torque fluctuation recognition module acquires the hip and knee joint trajectory during the take-off phase of competitive jumping, divides the continuous action segments, extracts the push-off connection link, determines whether the torque action switches states at the connection point, and obtains the joint action path adjustment reference segment. The path mutation switching module analyzes the change in the trajectory direction of the supporting leg based on the joint motion path adjustment reference segment, identifies the offset point, extracts the front and rear trajectory segments and adjusts the connection order to obtain the joint trajectory compensation segment sequence structure. The impact interference identification module extracts the axial path during the landing phase based on the joint trajectory compensation segment sequence structure, compares the node attitude changes, identifies state change segments, and obtains a set of impact action structure interference segments. Based on the set of impact action structure interference segments, the traction control adjustment module extracts the knee and ankle path segments, locates the traction start and end points, analyzes the action sequence deviation, and obtains the traction action response sequence structure interval. Based on the traction action response sequence structure interval and compensation segment sequence structure, the trajectory compensation control integration module advances the process of connecting each trajectory segment according to the starting point, execution point and transition point, matching it into a coherent control path, and obtaining the trajectory compensation control structure set.

[0006] As a further aspect of the present invention, the joint motion path adjustment reference segment includes the take-off preparation motion starting path, the push-off transition node path, and the motion transition switching path; the joint trajectory compensation segment sequence structure includes the supporting leg directional offset point, the trajectory segments on both sides of the directional offset point, and the adjusted trajectory fragment; the impact motion structure interference fragment set includes motion state change points, impact feature segments within the candidate range, and joint axial abnormal nodes; the traction motion response sequence structure interval includes the traction motion starting point, the motion ending position, and the motion sequence delay segment; and the trajectory compensation control structure set includes trajectory segment connection nodes, the starting execution point and the target node, and the path transition structure segment.

[0007] As a further aspect of the present invention, the push-off and extension connection link refers to the joint movement transition section between knee flexion and push-off in the jumping action, analyzing the changes in joint torque and angular velocity, identifying the start and end points of the stage, and determining whether there is a state switch. The axial path during the landing phase refers to the movement trajectory of the joints in the axial direction when landing from a jump. By analyzing the changes in the spatial position and attitude of the nodes, impact interference characteristics can be identified.

[0008] As a further aspect of the present invention, the traction start and end points refer to the two nodes between the knee joint and the ankle joint that represent the start and end of the traction movement in the action path, and the sequence and coordination of the action response are analyzed. The compensation segment sequence structure refers to the set of joint trajectory compensation paths constructed by identifying, correcting, and splicing the directional change points and preceding and following trajectory segments in the original motion trajectory.

[0009] As a further aspect of the present invention, the torque fluctuation identification module includes: The motion trajectory extraction submodule obtains motion data of the take-off preparation phase in competitive jumping events, extracts the position information sequence of the hip and knee joints within the phase, arranges the joint path points in chronological order, divides the continuous motion segments, and obtains the joint motion phase division sequence. The connecting structure positioning submodule divides the sequence based on the joint action phases, extracts the boundary nodes of adjacent phases, locates the end points and start points between phases, filters the node groups that switch between knee flexion and extension in the take-off action, locates the corresponding action transition points through the direction change relationship, and obtains the set of connecting nodes for the extension action. The state switching judgment submodule extracts the torque direction and angular velocity data corresponding to the nodes based on the set of connection nodes of the push-off action, selects the nodes in the switching area to compare the opposing states, identifies the position segment where the action driving direction changes, and obtains the joint action path adjustment reference segment.

[0010] As a further aspect of the present invention, the path abrupt switching module includes: The direction continuity judgment submodule extracts the time series and spatial direction vector of the trajectory nodes based on the support leg movement trajectory in the joint movement path adjustment reference segment, judges the changing trend of the direction vector between adjacent nodes, filters out the position points with discontinuous direction changes, and analyzes the position segments in the trajectory to obtain the trajectory direction change positioning sequence. The offset node extraction submodule verifies the node index corresponding to the direction change based on the trajectory direction change positioning sequence, locates the front and back positions of the node in the original trajectory path, extracts the path segments connecting the two sides of the node, and obtains a set of trajectory offset path segments. The motion path correction submodule extracts the start point, end point and direction vector of each segment based on the set of trajectory offset path segments, adjusts the connection relationship of each segment according to the actual advancement sequence of motion execution, corrects the directional offset of adjacent trajectory segments, and obtains the joint trajectory compensation segment sequence structure.

[0011] As a further aspect of the present invention, the impact interference identification module includes: The joint path extraction submodule extracts the time segment corresponding to the landing phase based on the joint trajectory compensation segment sequence structure, collects the joint's axial motion path data within the time range, filters continuous spatial nodes in the path, and obtains the joint landing motion path set. The posture change detection submodule calls the joint landing action path set, extracts the posture vector of each node, extracts the posture combination value sequence of adjacent nodes, compares the continuous change of the sequence, calculates the rate of change of the direction of the node posture curve, extracts the node index segment of the direction change, and obtains the joint posture change node sequence. The impact segment recognition submodule extracts trajectory segments before and after key nodes based on the joint posture change node sequence, analyzes the consistency of the starting and ending directions of each action trajectory, determines whether there are direction reversal and motion interruption features, filters action segments with joint impact characteristics, and obtains a set of impact action structure interference segments.

[0012] As a further aspect of the present invention, the traction control adjustment module includes: The path segment extraction submodule extracts the joint trajectory path between the knee and ankle joints based on the impact action structure interference fragment set, locates the start and end nodes of the traction action within the impact segment, filters continuous motion nodes and corresponding time labels within the path segment, extracts continuous action data within the region, and obtains the traction action segment node sequence. The connection order judgment submodule calls the traction action segment node sequence, reads the node action execution order, judges whether the response order between nodes is consistent during the path advancement process, identifies the range of nodes with delayed start in adjacent actions, extracts the action segments covered by the range, and obtains the traction action response delay node interval. The action response coordination submodule extracts the action status of adjacent nodes based on the traction action response delay node interval, adjusts the connection relationship of the position segments with misaligned response timing, and repositions the order of the first and last nodes to obtain the traction action response sequence structure interval.

[0013] As a further aspect of the present invention, the process of extracting continuous action data within the path segment extraction submodule is as follows: based on the continuous motion nodes and corresponding time tags within the path segment, the continuous changes in the time interval between nodes are analyzed to filter the sequence of action nodes that continuously evolve over time. The process of identifying the range of nodes with delayed start-up in adjacent actions in the connection sequence determination submodule is as follows: by comparing the time order of nodes in the traction action segment node sequence during path advancement, the set of nodes lagging in response advancement is identified. The process of repositioning the order of the first and last nodes in the action response coordination submodule is as follows: the first and last nodes are repositioned in the sequence of nodes in the traction action segment according to the node time order, and the start and end points of the response process are divided according to the time progression order.

[0014] As a further aspect of the present invention, the trajectory compensation control integrated module includes: The trajectory segment access submodule, based on the trajectory segments in the traction action response sequence structure interval, calls the action nodes at the start and end points of each path segment, extracts the content of each segment in the joint trajectory compensation segment sequence structure, maps and compares the action node paths, and obtains the trajectory node sequence information set. The action node localization submodule reads the starting point, execution point and transition point of each action path based on the trajectory node sequence information set, determines whether there is a path sequence inconsistency between adjacent segments in the trajectory sequence, extracts the node index and motion direction mark in the corresponding sequence, and obtains the path sequence structure between action segments. The trajectory connection adjustment submodule compares the path relationship between the start and end nodes of the trajectory segment based on the execution point and transition point in the path sequence structure between the action segments, rearranges the sequence of actions of the misaligned and reversed segments, and obtains the trajectory compensation control structure set.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, the take-off action phase is divided by the change in joint torque state, a trajectory compensation structure is constructed by combining the directional offset points in the support path, impact interference segments are extracted based on the characteristics of abrupt attitude change, the action response process is adjusted according to the difference in the order of the start and end positions in the traction path, the connection process of trajectory segments is promoted by relying on the arrangement relationship of execution nodes, and a continuous path system covering multiple action transition nodes is constructed. This enhances the structural matching ability of the action to directional changes, impact disturbances and response delays, and improves the path extensibility, action coordination and connection stability of the joint trajectory under unsteady motion conditions. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a system flowchart of the present invention; Figure 2 This is a system block diagram of the present invention; Figure 3 This is a flowchart of the torque fluctuation identification module in this invention; Figure 4 This is a flowchart of the path mutation switching module in this invention; Figure 5 This is a flowchart of the impact interference identification module in this invention; Figure 6 This is a flowchart of the traction control adjustment module in this invention; Figure 7 This is a flowchart of the trajectory compensation control integrated module in this invention. Detailed Implementation

[0018] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0019] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0020] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0021] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0022] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0023] This invention provides an intelligent analysis device for dynamic compensation of joint trajectory in competitive sports, such as... Figure 1-2 The diagram shown illustrates an intelligent analysis device for joint trajectory dynamic compensation in competitive sports. The device includes: The torque fluctuation recognition module acquires the motion trajectory of the hip and knee joints during the take-off preparation phase in competitive jumping events. Based on the continuity of the motion, it divides the motion into stages, extracts the push-off and extension links in each stage, determines whether the joint torque changes state at the motion connection point, identifies the trajectory stage of the motion transition, and obtains the joint motion path adjustment reference segment. The path mutation switching module is based on the support leg motion trajectory in the joint motion path adjustment reference segment. It makes a continuity judgment on the directional change of each motion node and adjacent nodes, filters the position of directional offset in the connection, extracts trajectory segments on both sides of the directional offset point, adjusts the motion path before and after the offset, and obtains the joint trajectory compensation segment sequence structure. The impact interference identification module is based on the joint trajectory compensation segment sequence structure. It extracts the joint axial motion path during the landing phase, compares the node posture changes point by point along the path, identifies the position where the action state changes, sets the action segments before and after the position as candidate ranges, identifies action segments with joint impact transition characteristics within the candidate ranges, and obtains a set of impact action structure interference segments. The traction control adjustment module extracts the path segment between the knee and ankle joints based on the impact action structure interference fragment set, locates the start and end positions of the traction action within the segment, analyzes the action connection sequence, identifies the position segment where the traction response is delayed, and adjusts the action sequence within the path to keep the action in a coordinated state within a continuous range, thus obtaining the traction action response sequence structure interval. The trajectory compensation control integration module is based on the trajectory segments in the traction action response sequence structure interval, and connects the sequence of each segment in the joint trajectory compensation segment sequence structure. According to the arrangement relationship of the starting action point, target execution point and conversion node in the compensation path, it promotes the connection process between trajectory segments and obtains the trajectory compensation control structure set.

[0024] The joint motion path adjustment reference segment includes the starting path of the take-off preparation action, the path of the push-off transition node, and the action transition switching path. The joint trajectory compensation segment sequence structure includes the supporting leg directional offset point, the trajectory segments on both sides of the directional offset point, and the adjusted trajectory fragment. The impact action structure interference fragment set includes the action state change point, the impact feature segment within the candidate range, and the joint axial abnormal node. The traction action response sequence structure interval includes the traction action starting point, the action ending position, and the action sequence delay segment. The trajectory compensation control structure set includes the trajectory segment connection node, the starting execution point and the target node, and the path transition structure segment.

[0025] Specifically, such as Figure 2 , 3 As shown, the torque fluctuation identification module includes: The motion trajectory extraction submodule obtains motion data of the take-off preparation phase in competitive jumping events, extracts the position information sequence of the hip and knee joints within the phase, arranges the joint path points in chronological order, divides the continuous motion segments, and obtains the joint motion phase division sequence. To acquire motion data of the take-off preparation phase in competitive jumping events, the joint trajectory record files of athletes who have completed jumps in the corresponding scenarios are selected sequentially. The frame ranges covered by the push-off preparation motion are located within the record content. Data segments are extracted based on frame numbers and joint position information, and visual verification is performed using graphical trajectory diagrams. Based on this, the position information sequences of the hip and knee joints are extracted. During the extraction process, the joint coordinate indices of each frame of pose data are traversed to construct a continuous sequence of joint path points. Only the key coordinate dimensions representing motion transitions are retained in the position information dimension, discarding static redundant dimensions to obtain standard joint path lines. This process is then used to arrange the joint path points. The data is linearly connected according to the frame order in the original data to eliminate discrete noise interference caused by non-target frames. Then, the action interval is initially demarcated according to the magnitude of the change in the direction of the coordinate points. The continuous direction consistency judgment method is adopted. By accumulating the number of times the direction change rate of the path points is accumulated, the direction inflection point segments with prominent change density are screened out. The positions with obvious changes in direction are marked as segment intervals. Path points are extracted between every two continuous direction change segments to form an interval path segment. The segmented action intervals such as initial push-off, transition force, and assist transfer are divided in sequence. Then, the joint paths in each segment are independently encoded to obtain the joint action stage division sequence.

[0026] The connecting structure positioning submodule is based on the sequence of joint action stages, extracts the boundary nodes of adjacent stages, locates the end point and start point between stages, filters the node group that switches between knee flexion and extension in the take-off action, locates the corresponding action transition point through the direction change relationship, and obtains the set of connecting nodes for the extension action. First, the start and end node numbers of each stage path segment in the sequence are called. Each pair of adjacent stages is traversed, and the path point coordinates of the end node of the current stage and the start node of the next stage are extracted from each stage pair. The spatial position index of the extracted nodes is recorded for subsequent direction change judgment. Then, a direction vector group is constructed for the coordinate points of the boundary between two adjacent stages. The angle between the direction vector from the end point to the start point and the path vector at the end of the stage is judged. If the angle is between 60° and 120°, it is considered that there is a transition in movement direction. Stage transition segments without direction change can be eliminated by judging this threshold range. Next, the node trajectories in the direction change segments are analyzed. Combining the descending trend of the hip and knee points in the knee flexion stage trajectory and the upward movement trajectory of the points in the extension stage, node groups where the vertical direction continuously reverses are identified. The process involves screening, removing node groups with a time span exceeding five frames, and detecting sudden increases in velocity variation at the starting position. A threshold is set based on the velocity variation difference, with a critical velocity variation difference of 1.5 rad / s². Two frames are taken before and after the velocity abrupt change point for velocity difference calculation. If the threshold is met, the node group is retained as a suspicious switching segment. Each suspicious segment is numbered and recorded. All node groups that meet the criteria are then aggregated. By comparing the overall take-off time range of the joint trajectory, further screening is conducted for node groups that exhibit a continuous contraction trend in the initial trajectory and a continuous force-propelling trend in the subsequent trajectory during the take-off phase. Finally, boundary point groups that satisfy the knee flexion and extension switching behavior characteristics are obtained, and their numbered segments are merged to form a unified node set, resulting in the set of nodes connecting the extension action.

[0027] The state switching judgment submodule is based on the set of nodes connecting the push-off action. It extracts the torque direction and angular velocity data corresponding to the nodes, selects the nodes in the switching area to compare the opposing states, identifies the position segment where the action driving direction changes, and obtains the joint action path adjustment reference segment. First, the torque direction information and angular velocity sequence corresponding to each node are extracted from the set. The torque direction is represented as the vector change trend along the main motion axis, and the angular velocity is represented as the rate of change of the joint rotation angle per unit time. By constructing two corresponding arrays, it is ensured that each node can correspond to a set of torque direction values ​​and angular velocity values. Then, two consecutive nodes in the node set are paired. A node pair list is constructed by using a sliding window traversal method. Subsequently, the torque direction values ​​in each node pair are compared for direction consistency. The judgment method is to determine the sign by calculating the angle between the vectors. If the angle between the torque direction vectors of two consecutive nodes is 175 degrees, it is considered to be a direction reversal state. Then, the angular velocity value sequence is judged for direction change trend. If the signs of the consecutive angular velocities are opposite, it is further confirmed. When the motion state changes, both of the above conditions must be met simultaneously to identify that the node pair is in a state transition behavior. Then, the node pairs that meet the above conditions are marked with their positions and their corresponding path segments are extracted as candidate state transition segments. After traversing all node pairs, a preliminary set of path segments is formed. Then, stable and continuous direction segments after the change of motion direction are identified in the set of path segments. The duration of the change of direction in the first half and the second half is counted in frames. It is required that the duration of both segments is more than five frames. If it is not met, the corresponding path segment is removed. After completion, the remaining path segments are further processed to exclude path segments with a continuous interval of less than five frames on the time axis and renumber them. Finally, a set of path position segments with clear changes in motion direction and prominent angular velocity turning are selected to obtain the joint motion path adjustment reference segment.

[0028] Specifically, such as Figure 2 , 4 As shown, the path abrupt change switching module includes: The direction continuity judgment submodule extracts the time series and spatial direction vector of the trajectory nodes based on the support leg motion trajectory in the joint motion path adjustment reference segment, judges the changing trend of the direction vector between adjacent nodes, filters out the position points with discontinuous direction changes, and analyzes the position segments in the trajectory to obtain the trajectory direction change positioning sequence. First, a traversal operation is performed on the included path segments, extracting time series information and spatial direction vectors for each node. The time label of each node is used as the sorting basis to distinguish the rhythmic order of the trajectory in continuous spatial motion. At the same time, its direction vector value in the three-dimensional coordinate system is recorded. A direction vector pair is formed for each pair of adjacent nodes, and the consistency of direction is determined based on the angle relationship between the two vectors. The size of the angle is indirectly determined by the vector dot product relationship without introducing a specific formula, only judging the trend of vector change. For example, if the angle between the direction vectors of nodes numbered 15 and 16 is between 175 degrees and 185 degrees, it can be considered that a direction reversal has occurred. Such node pairs are marked as direction jump points. After filtering out points with abrupt change characteristics from all node pairs, two frames are extended forward and backward from the abrupt change point as the center to extract the information containing... For the trajectory segment of the mutation point, a candidate segment set is constructed. The directional differences between adjacent start and end points of the candidate segments are compared to determine whether the change is continuous. If the directional consistency score between two adjacent mutation segments is lower than the preset directional consistency threshold, for example, the directional consistency rate is set to be lower than 0.3, then the segment can be marked as a discontinuous trajectory segment. The positions of all discontinuous segments on the trajectory time axis are recorded and sorted synchronously to ensure that their position in the overall trajectory can be restored according to the motion sequence. Combined with actual examples, such as in a certain jump action, there are segments numbered 24 to 30 in the support leg path that have obvious directional changes, deflecting backward between 24 and 25, and deflecting forward again between 28 and 29. The directional changes in the middle are not continuous. Based on the above judgment logic, this segment is included as an abnormal position point of directional change, and finally the trajectory direction change positioning sequence is obtained.

[0029] The offset node extraction submodule locates the sequence based on the trajectory direction change, verifies the node index corresponding to the direction change, locates the front and back positions of the node in the original trajectory path, extracts the path segments on both sides of the connecting node, and obtains the set of trajectory offset path segments. First, each marked direction change node is read one by one. Each node contains its index number in the support leg trajectory and its corresponding spatial direction data. For each node, a backtracking operation is performed on the index position to find its forward and backward adjacent points in the original trajectory sequence. The indices of the preceding and following nodes in the path are recorded respectively, and the path boundary range of the segment where the node is located is constructed. The forward range is set to extend forward by four nodes, and the backward range is set to extend backward by four nodes, forming a set of path segments centered on the direction change node. Then, the positions of the nodes in the path segments are checked. If consecutive nodes... If the path vector difference exceeds the set directional fluctuation threshold, the segment is identified as a potential offset segment. For example, in a jump test, an athlete's supporting leg trajectory shows directional fluctuation points between node numbers 38 and 46, with node number 41 being a directional change positioning point. The path segment numbered 37 to 45 is extracted, the vector difference value of each pair of consecutive nodes in the path segment is recorded, and the position sequence with frequent directional changes is screened out. This segment is then included as a trajectory offset segment. Multiple directional fluctuation segments are then numbered sequentially and compared synchronously with the original trajectory path to confirm their actual position in the action, ultimately obtaining the set of trajectory offset path segments.

[0030] The motion path correction submodule extracts the start point, end point and direction vector of each segment based on the set of trajectory offset path segments, adjusts the connection relationship of each segment according to the actual advancement sequence of motion execution, corrects the directional offset of adjacent trajectory segments, and obtains the joint trajectory compensation segment sequence structure. First, extract the coordinates of the first and last nodes of each path segment in the set as the start and end points, and obtain the direction vector between the nodes. The direction vector reflects the dominant movement direction of the path segment. For example, in the jump action, the offset segment numbered P23 to P31 has the coordinates of the start point P23 and the end point P31 as (2.1, 3.5, 1.7) and (5.4, 6.2, 3.3) respectively. Based on this, extract the dominant direction vector of this segment as the coordinate difference vector. Then, using the time progression order as a reference, check the connection direction relationship between adjacent segments one by one to determine whether there is an angular deviation between the direction vectors. If there is... If a deviation occurs, the connection direction of the path segment is adjusted. Specifically, the adjustment action is to reverse the arrangement order of the nodes within the path segment or to perform an internal translation operation on the node order until the direction is continuous after the adjacent segments are connected. On this basis, the offset segment P23-P31 is selected and compared with the adjacent segment P32-P40. If the angle between their connection vectors exceeds the set connection offset reference threshold of 45 degrees, it is confirmed that the direction is discontinuous. After performing the direction alignment operation, the path connection is performed again. The operation is repeated until all offset segments are processed. After all the corrected path segments are sorted in order, the path is spliced ​​to obtain the joint trajectory compensation segment sequence structure.

[0031] Specifically, such as Figure 2 , 5As shown, the impact interference identification module includes: The joint path extraction submodule is based on the joint trajectory compensation segment sequence structure. It extracts the time segment corresponding to the landing phase, collects the joint's motion path data in the axial direction within the time range, filters continuous spatial nodes in the path, and obtains the joint landing motion path set. First, extract the time segment corresponding to the landing phase of the action from the structure. Specifically, read the start and end index numbers marked as landing cycles in the compensation segment sequence and locate the time range. For example, if the action segment numbered T128 to T165 is defined as the landing phase, then extract all joint path data within this time interval. Then, collect the continuous action coordinate points of the joints along the axial direction within this time segment range and organize them according to the sampling order. Next, call the node index number to compare the spatial distance between each node one by one, and sequentially filter out the data nodes identified as breakpoints in the continuity judgment. During the filtering process, the displacement between nodes is used as the basis for the filtering. A distance less than 0.8 meters from the set spatial continuity benchmark is used as the continuous node identification standard. For any pair of nodes Pi and Pi+1, if their Euclidean distance is less than the benchmark value, they are retained as valid continuous nodes; otherwise, they are discarded. This filtering operation is repeated in multiple compensation segments, and the filtered data points are classified and organized using the joint number as the index. For example, in the knee joint path numbered J3, nodes P135 to P148 form a continuous segment, which meets the filtering standard and is retained. Finally, a set of data constrained by both time segments and spatial continuity is formed on each joint, resulting in the joint landing action path set.

[0032] The posture change detection submodule calls the joint landing motion path set, extracts the posture vector of each node, extracts the posture combination value sequence of adjacent nodes, compares the continuous change of the sequence, calculates the rate of change of the direction of the node posture curve, extracts the node index segment of the direction change, and obtains the joint posture change node sequence. The specific formula for calculating the rate of change of the direction of the nodal attitude curve is as follows: ; in, Representing the The rate of change of the orientation of each node The set of valid frame sequences representing a node. Represents the frame time interval. Representing the The node at the th The attitude and orientation angle of the frame. Represents the average value of the direction angle. The representative node is at the Frame pose vector magnitude, Represents the average modulus length. The representative node is at the The squared value of the two-dimensional positional movement distance on the frame. Represents a positive real constant that avoids division by zero. Represents the modulus length difference adjustment coefficient; node The following actual collected data was recorded in 5 frames: Attitude direction angle sequence (unit: rad): ; Attitude vector magnitude sequence (unit: m): ; Coordinate point sequence: ; The sequence of squared displacement values ​​(unit: m²) calculated from the coordinates is as follows: Frame 2 ; Frame 3 ; Frame 4 ; Frame 5 ; Calculate the average orientation angle and average modulus: , ; Substituting each frame into the formula to calculate the five values, and then summing them up, we get: sum ; Substitute into the formula: ; This result indicates that the node The attitude change rate within the current time period is 15.652 rad / s, which is considered a sudden change in motion. If a preset threshold for the direction change rate is set... Then the current node satisfies Proceed to the node sequence extraction process; The advantage of the formula is that by introducing a multi-parameter hybrid mechanism of modulus deviation, coordinate displacement square and orientation angle deviation, it can uniformly measure the inconsistency of angular amplitude, spatial change amplitude and intensity in motion mutation, filter out invalid fluctuations, and improve the stability of attitude change detection.

[0033] The impact fragment recognition submodule extracts trajectory segments before and after key nodes based on the joint posture change node sequence, analyzes the consistency of the starting and ending directions of each action trajectory, determines whether there are direction reversal and motion interruption features, filters action segments with joint impact characteristics, and obtains a set of impact action structure interference fragments. First, the trajectory segment corresponding to each key node index is extracted. Several path points before and after each node are used as the analysis range. For example, for the key node numbered N58, five nodes P53 to P57 are extracted forward, and five nodes P59 to P63 are extracted backward, forming trajectory segments P53 to P63 as a group of analysis units. Then, in each trajectory segment, the unit vectors of the starting and ending directions are identified, and the angle between the vectors is recorded as a basis for directional consistency. If the angle between the starting and ending directions exceeds the set range, with an angle change of 170 degrees, the segment is marked as having a direction reversal feature. Subsequently, the [missing information - likely a specific detection method or function] is then performed. If the velocity change sequence of nodes within a segment contains multiple intervals where the velocity approaches zero and is accompanied by abrupt velocity changes, such as a sudden drop in velocity from 0.12s to 0.01s at node P57 and then a rapid rebound to 0.15s, it indicates a sign of motion interruption at that location. This segment of trajectory is retained as a candidate segment with impact characteristics. Then, the joint numbers in multiple candidate segments are uniformly verified, and segments with mismatched action types are eliminated. Only the trajectory with the action type and number is retained. For example, the knee joint trajectory segment with the number J2 is selected as a qualified result, and finally, the set of impact action structure interference segments is obtained.

[0034] Specifically, such as Figure 2 , 6 As shown, the traction control adjustment module includes: The path segment extraction submodule extracts the joint trajectory path between the knee and ankle joints based on the impact action structure interference fragment set, locates the start and end nodes of the traction action within the impact segment, filters the continuous motion nodes and corresponding time labels within the path segment, extracts the continuous action data within the region, and obtains the traction action segment node sequence. First, the motion node numbers corresponding to the knee and ankle joints in the trajectory sequence are extracted. For example, in segment number C12, the knee joint motion path ranges from K33 to K50, and the ankle joint motion path ranges from A31 to A48. The path interval between them constitutes the analysis range. Next, the motion trajectory time labels within this range are identified, and the timestamp of each node is extracted as the sequence T33 to T48. Then, combined with the spatial displacement direction of each pair of joints, the positions of continuous nodes on the motion path are determined. For any pair of adjacent nodes, if their velocity change is less than the threshold of 0.05 m / s and the displacement direction deviation does not exceed 10 degrees, then this pair of nodes is considered... The nodes that meet the continuity criteria are identified as continuous motion points and are then linked together to form a path segment identification sequence, such as P1, P2, P3, etc. Based on the time tags contained in the path segment sequence, the start and end nodes of the traction action are located. For example, if the P1 path segment starts at T34 and ends at T44, then the start node is K34 and the end node is K44. Finally, within the start and end range, the motion displacement value, direction vector, and attitude angle value are extracted for each node. Interruption points and duplicate points are removed, and continuously changing data segments are retained to form a set of nodes with complete motion direction and time index, ultimately obtaining the traction action segment node sequence.

[0035] The connection order judgment submodule calls the traction action segment node sequence, reads the node action execution order, judges whether the response order between nodes is consistent during the path advancement process, identifies the range of nodes with delayed start in adjacent actions, extracts the action segments covered by the range, and obtains the traction action response delay node interval. First, the sorting information of each node in the time index is read. The nodes are arranged according to the timestamp sequence T1 to Tn when the action is collected, forming the node execution order set S. Each element in S corresponds to a spatial location and execution time. Then, the progress of each pair of adjacent nodes in the time sequence is judged, and node pairs with abnormal responses are extracted, that is, the displacement start time of the later node exceeds the last action time of the earlier node, which exceeds the delay threshold. The threshold is set with reference to 1.5 times the average action interval of the landing phase. For example, when the average time difference of joint advancement is 0.12 seconds, the delay threshold is set to 0.18 seconds. Based on this, the displacement of each node in the node sequence is counted. The start and end times are calculated, and the time difference between node pairs is determined. If the start time of node Bi is 1.84 seconds and the last execution time of node Ai is 1.63 seconds, the difference between the two is 0.21 seconds, which exceeds the set threshold. Therefore, Bi is identified as a response delay node. Then, the range of nodes with continuous delays is selected from the entire sequence. The path interval between the start delay node and the end delay node is extracted as the response anomaly segment. Then, combined with the spatial path and action direction label corresponding to each node, an action segment set containing time and spatial attributes is constructed. Finally, this segment set is constructed into a dual index structure of time interval and node index to obtain the traction action response delay node interval.

[0036] The action response coordination submodule extracts the action status of adjacent nodes based on the traction action response delay node interval, adjusts the connection relationship of the position segments where the response timing is misaligned, and repositions the order of the first and last nodes to obtain the traction action response sequence structure interval. First, the action labels and time indices corresponding to each node in the interval are read to identify the motion stage of each node when the action state changes. By comparing the continuity of spatial position and the relative position of time labels between each pair of adjacent nodes, node pairs with inconsistent path progression order are extracted to determine whether there is a response offset between nodes. For example, if the start time of node A is marked as 2.35 seconds, its spatial position is earlier than that of node B on the path, while the start time of node B is 2.18 seconds, this is identified as a response misalignment. Based on this, the path segment of such node pairs is set as a segment to be processed. Then, the connection order between nodes in this path segment is adjusted, and the time labels are rearranged according to the spatial progression order to ensure that the node time values ​​in the path progression direction are monotonically increasing. Then, the first and last nodes in the adjusted node set are located, their corresponding action states are re-extracted, and compared with the state before adjustment. If there is a change in stage label, the state type of the node is corrected. After the sequence correction of all action segments is completed, the continuity of the path segments is re-marked according to the node index order, thereby establishing a consistent progression relationship in both time and space, and finally obtaining the traction action response sequence structure interval.

[0037] Specifically, such as Figure 2 , 7 As shown, the trajectory compensation control integrated module includes: The trajectory segment access submodule is based on the trajectory segments in the traction action response sequence structure interval. It calls the action nodes at the start and end points of each path segment, extracts the content of each segment in the joint trajectory compensation segment sequence structure, maps and compares the action node paths, and obtains the trajectory node sequence information set. First, identify the start and end points of each path segment. Then, retrieve the node numbers and spatial coordinates of each action node in the node sequence, read the action type identifier of each node, and use the node index value to locate its position in the original data. Next, extract the first and last nodes within each path segment. From the joint trajectory compensation segment sequence structure, sequentially compare the path segments located by the node indexes to confirm whether there is a start node in the compensation path that matches the action point in the current trajectory segment. The matching of path segments is confirmed by comparing node labels with spatial vectors. For example, if the start point number of the current trajectory segment is n1 and the end point number is n2, and there are... In a path segment with a starting point numbered n1 but an ending point numbered n3, the path segment is considered mismatched. It is necessary to continue searching for a segment that matches n2 in the subsequent segments and determine whether the direction of action is consistent through the node index sequence. If the starting node index is less than the ending node index, the direction is normal; otherwise, the order needs to be readjusted. After confirming the matching relationship between the starting and ending points and the content of the compensation segment in each trajectory segment, the successfully matched node segment numbers are extracted into a group to form a path index mapping table corresponding to the trajectory segment. Through the index relationship in the table, each action node is bound to its corresponding trajectory path, and finally, the trajectory node sequence information set is obtained.

[0038] The action node localization submodule reads the starting point, execution point and transition point of each action path based on the trajectory node sequence information set, determines whether there is a path sequence inconsistency between adjacent segments in the trajectory sequence, extracts the node index and motion direction mark in the corresponding sequence, and obtains the path sequence structure between action segments. First, the starting point, execution point, and transition point marked in each action path are read. The node number and its corresponding action label are obtained. The direction vector formed by the starting and ending points of each path is compared. The node index arrangement of the current segment and adjacent segments is read from the sequence information set. It is determined whether the connection between the starting and ending points follows an ascending order of node numbers. If the ending point number of the previous segment is greater than the starting point number of the next segment, it is marked as an inconsistent path order. Then, the direction identifier within the corresponding segment is extracted through the action path label. For example, if the direction identifier of the previous segment is "forward advance" and the next segment is "backward retreat," it is marked as a reversal on the logical path. The node time label is used to determine whether the direction reversal is accompanied by a time misalignment. If the starting node of the next segment... If the timestamp is earlier than the timestamp of the previous termination node, it is further identified as a response conflict node segment. During the processing, three types of key points are extracted for each path segment: the first node at the beginning of the path, the point of maximum angular velocity change in the middle, and the termination node at the end of the path. The index values ​​of the above nodes in the original action sequence are used as the path segment feature vector for subsequent sorting and reconstruction processing. Then, the set of path segment numbers with inconsistent order are recorded as abnormal paths, and all node indices contained in them are extracted. The node index values ​​are rearranged from smallest to largest and a mapping table is established. At the same time, the direction identifier of each node is extracted, and the path segment is divided based on the direction identifier as the sequence component. The point where the direction change occurs is stored as the transition point index, and finally the path sequence structure between action segments is obtained.

[0039] The trajectory connection adjustment submodule compares the path relationship between the start and end nodes of the trajectory segment based on the execution point and transition point in the path sequence structure between action segments, rearranges the action sequence of the misaligned and reversed segments, and obtains the trajectory compensation control structure set. First, the starting and ending node indices of each path segment are read and their positions in the overall trajectory sequence. The time stamps between the starting and ending nodes are compared to determine if the time progression is consistent with the node number progression. If a path has ascending node numbers but reversed timestamps, it is identified as a misaligned segment, and the start and end positions of the corresponding segment are marked. Then, the segment and its adjacent path segments are labeled with direction vectors, and the direction change data from the start to the end point is extracted. The direction of the end point of the previous segment is calculated with the direction of the start point of the next segment to determine if there is a direction reversal. If the dot product of the direction vectors is less than 0.2, it is marked as a reverse segment. Finally, the segment numbers corresponding to all misaligned and reverse segments are summarized, and their full positions are extracted according to their positions in the original path sequence. The system indexes nodes and sorts them by timestamp in ascending order to generate a new sequential sequence for path segment reordering. During this process, the direction identifier of each segment needs to be read and paired with adjacent segments. If the direction identifiers of two adjacent segments are inconsistent, the end point of the previous segment is used as the starting point of the next segment, and the connection point of the two segments is updated. If the direction identifiers of two adjacent segments are consistent but there is a time sequence conflict, the smaller timestamp is used as the starting point and the larger timestamp as the ending point, and the direction vector and the start and end positions of the path are adjusted. If there are three or more consecutive segments with inconsistent directions during the reordering of multiple segments, a sliding window method is constructed to reorder them sequentially with adjacent segments as the processing unit, updating the node sequence and path vector data of each segment, and finally obtaining the trajectory compensation control structure set.

[0040] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An intelligent analysis device for dynamic compensation of joint trajectory in competitive sports, characterized in that, The device comprises: The torque fluctuation identification module obtains the hip and knee joint trajectory in the take-off stage of competitive jumping, divides the continuous action segment, extracts the stretch connection link, judges whether the torque action switches at the connection, and obtains the joint action path adjustment reference segment; The path mutation switching module analyzes the support leg trajectory direction change based on the joint action path adjustment reference segment, identifies the offset point, extracts the front and rear trajectory segments and adjusts the connection order, and obtains the joint trajectory compensation segment sequence structure; The impact interference identification module extracts the axial path in the landing stage based on the joint trajectory compensation segment sequence structure, compares the node posture change, identifies the state mutation segment, and obtains the impact action structure interference segment set; The traction control adjustment module extracts the knee and ankle path segment based on the impact action structure interference segment set, positions the traction start and stop points, analyzes the action sequence deviation, and obtains the traction action response sequence structure interval; The trajectory compensation control integrated module promotes each trajectory to connect according to the starting point, execution point and conversion point based on the traction action response sequence structure interval and compensation segment sequence structure, matches the continuous control path, and obtains the trajectory compensation control structure set.

2. The athletic sports-oriented joint trajectory dynamic compensation intelligent analysis device according to claim 1, characterized in that, The joint action path adjustment reference segment includes the take-off preparation action starting path, the stretch connection conversion node path, and the action conversion switching path.

3. The athletic sports-oriented joint trajectory dynamic compensation intelligent analysis device according to claim 1, characterized in that, The stretch connection link refers to the joint motion transition section between knee flexion and stretch in the jumping action, the change of joint torque and angular velocity is analyzed, the start and end nodes of the stage are identified, and whether there is state switching is judged. The axial path in the landing stage refers to the action trajectory path of the joint in the axial direction when jumping, and the impact interference characteristics are identified through the analysis of the change of node space position and posture.

4. The intelligent analysis device for dynamic compensation of joint trajectory in competitive sports according to claim 1, characterized in that, The traction start and stop points refer to the two nodes representing the start and end of traction motion in the action path between the knee joint and the ankle joint, and the sequence and coordination of the action response are analyzed. The compensation segment sequence structure refers to the joint trajectory compensation path set constructed by identifying, correcting and splicing the direction mutation point and the front and rear trajectory segments in the original action trajectory.

5. The intelligent analytical device for dynamic compensation of joint trajectory in competitive sports according to claim 1, characterized in that, The torque fluctuation identification module comprises: The action trajectory extraction submodule obtains the action data in the take-off preparation stage of competitive jumping events, extracts the position information sequence of the hip joint and the knee joint in the stage, arranges the joint path points in time sequence, divides the continuous motion action segment, and obtains the joint action stage division sequence; The connection structure positioning sub-module is configured to divide the joint action stage sequence, extract boundary nodes of adjacent stages, locate end points and start points between stages, screen node groups of knee bending and knee extension switching in take-off actions, locate corresponding action transition points through direction change relations, and obtain a knee extension action connection node set; The state switching judgment sub-module is configured to extract torque direction and angular velocity data corresponding to the nodes based on the knee extension action connection node set, select nodes in a switching region for comparison, identify positions of action driving direction changes, and obtain a joint action path adjustment reference segment.

6. The intelligent analytical device for dynamic compensation of joint trajectory in competitive sports according to claim 1, characterized in that, The path mutation switching module includes: The direction continuity judgment sub-module is configured to extract time sequences and spatial direction vectors of trajectory nodes based on support leg movement trajectories in the joint action path adjustment reference segment, judge change trends of direction vectors between adjacent nodes, screen position points with discontinuous direction changes, analyze position segments in the trajectories, and obtain a trajectory direction change positioning sequence; The offset node extraction sub-module is configured to verify node indexes corresponding to direction changes based on the trajectory direction change positioning sequence, locate positions of the nodes before and after the nodes in original trajectory paths, extract path segments on both sides of the nodes, and obtain a trajectory offset path segment set; The action path correction sub-module is configured to extract start points, end points, and direction vectors in each segment based on the trajectory offset path segment set, adjust connection relations of each segment according to actual pushing sequences of action execution, correct direction offsets of adjacent trajectory segments, and obtain a joint trajectory compensation segment sequence structure.

7. The intelligent analytical device for dynamic compensation of joint trajectory in competitive sports according to claim 1, characterized in that, The impact interference identification module includes: The joint path extraction sub-module is configured to extract time segments corresponding to landing stages based on the joint trajectory compensation segment sequence structure, collect joint action path data in an axial direction within a time range, screen continuous spatial nodes in the path, and obtain a joint landing action path set; The posture change detection sub-module is configured to call the joint landing action path set, extract posture vectors of each node, extract posture combination value sequences of adjacent nodes, compare continuous change conditions of the sequences, calculate direction change rates of node posture curves, extract node index segments of direction changes, and obtain a joint posture change node sequence; The impact segment identification sub-module is configured to extract trajectory segments before and after key nodes based on the joint posture change node sequence, analyze action start point and end point direction consistency of each action trajectory, judge whether there are direction flipping and motion interruption features, screen action segments with joint impact features, and obtain an impact action structure interference segment set.

8. The intelligent analytical device for dynamic compensation of joint trajectory in competitive sports according to claim 1, characterized in that, The traction control adjustment module includes: The path segment extraction sub-module is configured to extract joint trajectory paths between knee joints and ankle joints based on the impact action structure interference segment set, locate start and end nodes of traction actions in impact segments, screen continuous motion nodes and corresponding time labels in path segments, extract action continuous data in the region, and obtain a traction action segment node sequence; The connection sequence judgment submodule calls the traction action section node sequence, reads the node action execution sequence, judges whether the response sequence of the nodes in the path promotion process is consistent, identifies the node range of the delayed start in the adjacent actions, extracts the action fragments covered by the range, and obtains the traction action response delay node interval; The action response coordination submodule extracts the adjacent node action state based on the traction action response delay node interval, adjusts the position segment connection relationship of the response time sequence misalignment, repositions the first and last node order, and obtains the traction action response sequence structure interval.

9. The intelligent analysis device for dynamic compensation of joint trajectory in competitive sports according to claim 8, characterized in that, The process of extracting the action continuous data in the extraction area in the path section extraction submodule is specifically: according to the continuous motion nodes in the path section and the corresponding time label, the continuous change of the time interval between the nodes is analyzed, and the action node sequence continuously evolved in the time promotion is screened; The process of identifying the node range of the delayed start in the adjacent actions in the connection sequence judgment submodule is specifically: by comparing the time sequence of the nodes in the traction action section node sequence in the path promotion, the set of nodes delayed in the response promotion is identified; The process of repositioning the first and last node order in the action response coordination submodule is specifically: according to the node time sequence, the positions of the first node and the last node in the traction action section node sequence are re-set, and the start point and the end point of the response process are divided according to the time promotion sequence.

10. The intelligent analytical device for dynamic compensation of joint trajectory in competitive sports according to claim 1, characterized in that, The trajectory compensation control integrated module comprises: The trajectory fragment access submodule extracts the content of each segment in the joint trajectory compensation segment sequence structure based on the action nodes corresponding to the starting point and the termination point of each path in the traction action response sequence structure interval, maps and compares the action node path, and obtains the trajectory node sequence information set; The action node positioning submodule reads the starting point, the execution point and the conversion point of each action path based on the trajectory node sequence information set, judges whether there is a path sequence inconsistency between adjacent paragraphs in the trajectory sequence, extracts the node index and the motion direction mark in the corresponding sequence, and obtains the action segment path sequence structure; The trajectory connection adjustment submodule compares the path relationship between the trajectory segment starting and termination nodes based on the execution point and the conversion point in the action segment path sequence structure, rearranges the paragraph action order of the path misalignment and the reverse, and obtains the trajectory compensation control structure set.

Images