A somatosensory interaction-based human health state monitoring and analysis method

Abstract

The application discloses a kind of based on somatosensory interaction human health state monitoring and analysis method, comprising the following steps: standard somatosensory interaction data is generated by collecting data;Standard somatosensory interaction data is generated based on somatosensory task action data;The action contribution of target body part and non-target body part to task completion process is calculated respectively, and body part action contribution sequence is generated;Identify the action process that target body part action contribution drops, non-target body part action contribution rises and task completion state keeps complete, and generate action compensation event;Action compensation migration path is generated, and action compensation load data is calculated;Compensation failure somatosensory interaction task is generated, and action response during acquisition generates compensation failure response data;Local motor ability decline is identified, and human health state monitoring and analysis result is generated.The application uses somatosensory action contribution and compensation migration analysis, identifies hidden local motor ability decline, improves monitoring accuracy.

Description

Technical Field

[0001] This invention relates to the field of motion-sensing interaction technology, and in particular to a method for monitoring and analyzing human health status based on motion-sensing interaction. Background Technology

[0002] Human health monitoring methods typically utilize somatosensory devices to collect information on human position, joint angles, and task completion. The monitoring assesses the subject's motor abilities by measuring movement range, trajectory, completion time, and interaction target triggering. Some methods can also identify compensatory phenomena involving the trunk, shoulder, or unaffected limbs in task movements, thereby aiding in determining the correctness of movement technique.

[0003] However, relying primarily on task completion or motion data from individual body parts makes it difficult to quantify the actual contributions of target and non-target body parts to task completion, and also hinders the identification of the transfer of motion contributions between different body parts. When non-target body parts maintain task completion through continuous compensation, the decline in local motor ability of the target body parts can be easily masked, leading to insufficient accuracy in the analysis of human health status.

[0004] Therefore, how to provide a method for monitoring and analyzing human health status based on somatosensory interaction is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] One objective of this invention is to propose a method for monitoring and analyzing human health status based on somatosensory interaction. This invention utilizes somatosensory movement contribution and compensatory migration analysis to identify hidden local declines in motor ability and improve monitoring accuracy.

[0006] A method for monitoring and analyzing human health status based on somatosensory interaction according to an embodiment of the present invention includes the following steps: Collect raw somatosensory interaction data during the initial somatosensory interaction task performed by the monitored object, synchronize the somatosensory interaction data in time, unify the coordinates and associate the raw somatosensory interaction data with body parts, and generate standard somatosensory interaction data; Based on standard motion-sensing interaction data, human motion processes are divided into target body parts and non-target body parts to generate motion-sensing task motion data. Based on the motion data of the motion-sensing task and the target trajectory corresponding to the initial motion-sensing interaction task, the motion contribution of the target body parts and non-target body parts to the task completion process is calculated respectively, and a sequence of body part motion contribution is generated. Based on the body part motion contribution sequence and somatosensory task motion data, identify the motion process in which the motion contribution of the target body part decreases, the motion contribution of the non-target body part increases, and the task completion status remains complete, and generate motion compensation events. Based on the motion compensation events, motion compensation migration paths are generated, and the additional motion, compensation duration, and number of repetitions for each non-target body part are calculated to generate motion compensation load data. Based on the motion compensation migration path, the task structure of the initial somatosensory interaction task is changed to generate a somatosensory interaction task with compensation failure, and the motion response during the collection period is used to generate compensation failure response data. Based on motion compensation events, motion compensation migration paths, motion compensation load data, and compensation failure response data, the system identifies localized declines in motor ability masked by motion compensation, and generates human health status monitoring and analysis results.

[0007] Optionally, the raw data for the haptic interaction includes human motion data and task execution data; The raw motion-sensing interaction data is synchronized in time, unified in coordinates, and associated with body parts to generate standard motion-sensing interaction data.

[0008] Optionally, the generation of the motion-sensing task motion data includes: Obtain body part identifiers, position data, joint angle data, and task execution status corresponding to each sampling moment from standard somatosensory interaction data; In the initial somatosensory interaction task, the body part that is designated to perform the task action is identified as the target body part, and the body parts that undergo positional or joint angle changes during human movement, other than the target body parts, are identified as non-target body parts. The target body part's motion sequence is generated by connecting the position data and joint angle data corresponding to the target body part according to the sampling time, and the non-target body part's motion sequence is generated by connecting the position data and joint angle data corresponding to each non-target body part according to the body part identifier. The task execution status is connected according to the sampling time to generate a task completion sequence. The target body part action sequence, the non-target body part action sequence and the task completion sequence are associated according to the sampling time to generate motion-sensing task action data.

[0009] Optionally, the generation of the body part action contribution sequence includes: Align the sequence of actions of the target body parts and the sequence of actions of non-target body parts with the trajectory of the mission target in time, divide the continuous action intervals, and determine the target movement direction and target propulsion in each continuous action interval; The changes in position and joint angles of each body part within each continuous movement interval are obtained. The changes in position are projected onto the target movement direction, and the changes in joint angles are combined to generate the movement amount in the target direction. Based on the connection relationship between body parts, the motion quantities in each target direction are transferred to the target body part to generate the baseline task progress quantity. The motion quantities in each target direction corresponding to each body part are removed in turn, and the remaining motion quantities in the target direction are used to generate the culling task progress quantity. The difference between the baseline task advance and the elimination task advance is calculated. The difference is compared with the target advance to generate the action contribution value corresponding to each body part. The action contribution values ​​are then connected in the time sequence of continuous action intervals to generate the action contribution sequence of body parts.

[0010] Optionally, the generation of the action compensation event includes: The task completion status in the task completion sequence is mapped to the continuous action interval in the body part action contribution sequence. The change in action contribution of each body part between adjacent continuous action intervals is calculated to generate a body part action contribution change sequence. Within a continuous action interval where the task completion state remains complete, identify a continuous action interval where the action contribution of the target body part continuously decreases and the action contribution of at least one non-target body part continuously increases, and connect the continuous action intervals that are adjacent in time to generate candidate action compensation segments. According to the time sequence of continuous action intervals, the decrease in the action contribution of the target body part is matched with the increase in the action contribution of each non-target body part in turn to generate the action contribution acceptance amount of each non-target body part, and the non-target body part identifier and continuous action interval are associated to generate action contribution acceptance data. Based on the motion contribution reception data, non-target body parts that continuously generate motion contribution reception volume are identified. The target body parts, non-target body parts, candidate motion compensation segments, and motion contribution reception data are associated to generate motion compensation events.

[0011] Optionally, the generation of the action compensation load data includes: Data on target body parts, non-target body parts, and action contribution reception are obtained from action compensation events. Action compensation time sequence is generated according to the order in which each non-target body part first generates action contribution reception. Using the target body part as the starting point of migration, the first non-target body part is connected to the target body part. When the motion contribution of the previous non-target body part decreases and the motion contribution of the next non-target body part increases, and the corresponding continuous motion intervals of the two are connected, the previous non-target body part and the next non-target body part are connected to generate a motion compensation migration pair. Connect the target body parts and each action compensation transfer pair according to the action compensation sequence to generate the action compensation transfer path; For each non-target body part in the motion compensation migration path, the actual amount of motion during the motion contribution period and the baseline amount of motion within the same task phase are calculated. The difference between the two is determined as the additional amount of motion. Combining the time of generation of the motion contribution, the time of zeroing, and the number of occurrences in each motion compensation event, the compensation duration and the number of repeated participations are generated. The additional amount of motion, the compensation duration, and the number of repeated participations are associated with the corresponding non-target body parts to generate motion compensation load data.

[0012] Optionally, the generation of the compensatory failure response data includes: Obtain the triggering body parts, action triggering time periods, action triggering order, and task completion relationships corresponding to each interaction target in the initial somatosensory interaction task, associate them according to the interaction targets, and generate initial task structure data; According to the connection order from the target body part to each non-target body part in the action compensation migration path, the action contribution period corresponding to each non-target body part is matched with the action triggering period of each interactive target in the initial task structure data to determine the interactive target and action triggering period when each non-target body part participates in action compensation, and generate compensation path corresponding data. Process the data corresponding to the compensation path sequentially along the motion compensation migration path. According to the connection order of each non-target body part participating in motion compensation, obtain the interaction target corresponding to the adjacent non-target body part. Move the action triggering time of the interaction target corresponding to the next non-target body part forward to the action triggering time of the interaction target corresponding to the previous non-target body part that has not yet been completed, so that the action triggering time of the adjacent interaction targets overlaps with each other. Adjust the corresponding interaction targets from sequential triggering to simultaneous triggering. Regenerate the action triggering order according to the adjusted action triggering time and generate task structure adjustment data. The task structure adjustment data replaces the triggering body part, action triggering time period and action triggering order of the corresponding interactive target in the initial task structure data, while keeping the other interactive targets and task completion relationships unchanged, to generate a compensatory failure somatosensory interactive task. The system collects positional data, joint angle data, interactive target trigger status, and task completion status of the target body parts and non-target body parts during the execution of the compensatory failure somatosensory interaction task. These data are then connected according to the sampling time and body part identifier to generate the compensatory failure target body part action sequence, compensatory failure non-target body part action sequence, compensatory failure interactive target trigger sequence, and compensatory failure task completion sequence. These sequences are then correlated to generate compensatory failure response data.

[0013] Optionally, the generation of the human health status monitoring and analysis results includes: According to the target body part, the motion compensation events, motion compensation migration paths and motion compensation load data are matched, and the number of motion compensation events, the number of motion compensation migration paths, the number of non-target body parts, the extra movement volume of each non-target body part, the compensation duration and the number of repeated participations are counted to generate motion compensation feature data. The action sequences, interaction target trigger sequences, and task completion sequences are obtained from the compensation failure response data. The number of actions completed by the target body parts, the number of interaction targets triggered, and the task completion status are counted to generate compensation failure feature data. The continuous state of action compensation is identified based on action compensation feature data, and the failure state of compensation is identified based on compensation failure feature data. Based on the continuous state and failure state of the corresponding movement compensation for the target body part, the target body part that simultaneously has both states is identified as having a localized decline in motor ability that is masked by movement compensation. The corresponding movement compensation feature data and compensation failure feature data are then associated to generate human health status monitoring and analysis results.

[0014] The beneficial effects of this invention are: This invention collects human motion data and task execution data of the monitored object during the initial haptic interaction task, and synchronizes, unifies coordinates, and associates body parts. This allows for the analysis of target body parts, non-target body parts, and the task completion process within a unified temporal and spatial reference. Compared to existing methods that evaluate solely based on motion amplitude, task completion time, or interaction target triggering results, this invention can simultaneously reflect the participation of different body parts in the same task action, reducing analytical biases caused by data time misalignment, coordinate differences, and inaccurate body part correspondence, thus improving the completeness and consistency of the description of the human motion process.

[0015] This invention maps the positional changes and joint angle changes of each body part to the target motion direction corresponding to the task trajectory, calculates the motion contribution of both target and non-target body parts to the task propulsion process, and obtains the motion contribution value of each body part by removing the motion amount corresponding to the target direction of the body part to be analyzed one by one. Therefore, it no longer judges whether a body part participates in the task solely by whether it moves, but can quantify the actual propulsion effect of that body part on the task completion process, thereby distinguishing between normal coordinated movements and additional movements used to compensate for insufficient movement of the target body part.

[0016] This invention further correlates the reduction in motion contribution of the target body part with the increase in motion contribution of each non-target body part in a segmented manner to generate motion contribution transfer data. Based on the order and temporal connection of the motion contribution transfer amounts generated by different non-target body parts, a motion compensation migration path is formed. By calculating the additional motion amount, compensation duration, and number of repetitions for each non-target body part, the invention can simultaneously characterize the participating parts, migration direction, duration, and motion load of motion compensation, avoiding the problem of existing technologies that only identify a single compensation posture or count the number of compensations, thus failing to reflect a continuous compensation process.

[0017] This invention also adjusts the action triggering time and order of adjacent interactive targets in the initial somatosensory interaction task according to the action compensation migration path, so that the interactive targets that were originally triggered sequentially become a simultaneous triggering process, compressing the time connection process of different non-target body parts successively receiving action contributions, and generating a somatosensory interaction task with compensation failure. By combining the continuous state of action compensation in the initial task with the number of actions completed, the number of interactive targets triggered, and the task completion status in the somatosensory interaction task with compensation failure, it can identify the decline in local motor ability that is masked by action compensation under the results of conventional task completion, thereby improving the accuracy and pertinence of human health status monitoring and analysis results; Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is an overall flowchart of a human health status monitoring and analysis method based on somatosensory interaction proposed in this invention; Figure 2 This is a schematic diagram illustrating the adjustment of the somatosensory interaction task structure based on the motion compensation migration path in the somatosensory interaction-based human health status monitoring and analysis method proposed in this invention. Detailed Implementation

[0019] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0020] refer to Figures 1-2 A method for monitoring and analyzing human health status based on somatosensory interaction includes the following steps: Collect raw motion interaction data of the monitored object during the initial motion interaction task. The raw motion interaction data includes human motion data and task execution data. Synchronize the raw motion interaction data in time, unify coordinates, and associate body parts to generate standard motion interaction data. Based on standard somatosensory interaction data, human movement processes are divided into target body parts and non-target body parts to generate somatosensory task motion data. The somatosensory task motion data includes target body part motion sequences, non-target body part motion sequences, and task completion sequences. Based on the motion data of the motion-sensing task and the target trajectory corresponding to the initial motion-sensing interaction task, the motion contribution of the target body parts and non-target body parts to the task completion process is calculated respectively, and a sequence of body part motion contribution is generated. Based on the body part motion contribution sequence and somatosensory task motion data, identify the motion process in which the motion contribution of the target body part decreases, the motion contribution of the non-target body part increases, and the task completion status remains complete, and generate motion compensation events. Based on the motion compensation events, according to the temporal connection relationship between the body parts involved in motion compensation, a motion compensation migration path is generated, and the additional motion amount, compensation duration and repeated participation of each non-target body part are calculated to generate motion compensation load data. The task structure of the initial somatosensory interaction task is changed based on the motion compensation migration path to generate a compensated failure somatosensory interaction task. The motion response of the monitored object during the execution of the compensated failure somatosensory interaction task is collected to generate compensation failure response data. Based on motion compensation events, motion compensation migration paths, motion compensation load data, and compensation failure response data, the system identifies localized declines in motor ability masked by motion compensation, and generates human health status monitoring and analysis results.

[0021] In this embodiment, the raw data of the somatosensory interaction includes human motion data and task execution data; The raw motion-sensing interaction data is synchronized in time, unified in coordinates, and associated with body parts to generate standard motion-sensing interaction data.

[0022] In this embodiment, the generation of motion-sensing task motion data includes: Obtain body part identifiers, position data, joint angle data, and task execution status corresponding to each sampling moment from standard somatosensory interaction data; In the initial somatosensory interaction task, the body part that is designated to perform the task action is identified as the target body part, and the body parts that undergo positional or joint angle changes during human movement, other than the target body parts, are identified as non-target body parts. The target body part's motion sequence is generated by connecting the position data and joint angle data corresponding to the target body part according to the sampling time, and the non-target body part's motion sequence is generated by connecting the position data and joint angle data corresponding to each non-target body part according to the body part identifier. The task execution status is connected according to the sampling time to generate a task completion sequence. The target body part action sequence, the non-target body part action sequence and the task completion sequence are associated according to the sampling time to generate motion-sensing task action data.

[0023] In this embodiment, the generation of the body part movement contribution sequence includes: The motion sequence of the target body parts and the motion sequence of the non-target body parts in the motion data of the motion sensing task are time-aligned with the trajectory of the task target. The continuous motion intervals are divided according to adjacent sampling times, and the target motion direction and target propulsion amount corresponding to each continuous motion interval are determined. Specifically, both the standard somatosensory interaction data and the task target trajectory are recorded at 20 sampling times per second. The adjacent trajectory points with timestamps of 2100 ms and 2150 ms in the task target trajectory are corresponding to a continuous action interval. The two trajectory points with coordinates of (320, 410, 180) mm and (360, 420, 190) mm are connected to generate the target motion vector. The three coordinate components of the target motion vector are divided by the spatial length of the target motion vector to generate the target motion direction. The spatial distance between the two trajectory points is calculated to obtain the target propulsion amount of 42 mm for the corresponding continuous action interval. The positional and joint angle changes of the target body part and each non-target body part are obtained in each continuous motion interval. The positional changes of each body part are projected onto the corresponding target motion direction, and the target direction motion amount is generated by combining the joint angle changes. Specifically, the position data and joint angle data of the target body part and each non-target body part at the start and end of the continuous movement interval are read. Combined with the corresponding limb length and the connection relationship of the body parts, the change in joint angle is converted into the theoretical position change vector of the end of the corresponding body part. The theoretical position change vector is corrected by the actual collected position change vector. The corrected position change vector is projected onto the target movement direction. The right hand, forearm, upper arm, shoulder and torso generate target direction movement amounts of 18 mm, 4 mm, 5 mm, 7 mm and 9 mm respectively. Based on the connection relationship between body parts, the target direction motion quantity corresponding to each body part in the same continuous action interval is transferred to the target body part according to the connection relationship between body parts to generate the baseline task progress quantity. The target body part and each non-target body part are identified as the body parts to be analyzed in sequence. The target direction motion quantity corresponding to the body parts to be analyzed is removed, and the elimination task progress quantity is regenerated based on the target direction motion quantity corresponding to the remaining body parts. Specifically, following the connection sequence between the torso, shoulder, upper arm, forearm, and right hand, the target direction motion quantities corresponding to each body part within the same continuous motion interval are mapped to the right hand end. The positional change of the right hand end along the target motion direction is regenerated, resulting in a baseline task propulsion of 43 mm. The target direction motion quantity corresponding to the torso is set to zero, while keeping the target direction motion quantities corresponding to the shoulder, upper arm, forearm, and right hand unchanged. The positional change of the right hand end along the target motion direction is regenerated according to the same connection sequence, resulting in a culling task propulsion of 34 mm. Calculate the difference in propulsion between the baseline task propulsion and the elimination task propulsion, compare the propulsion difference with the target propulsion in the corresponding continuous action interval, and generate the action contribution value of the body part to be analyzed in the corresponding continuous action interval. The difference in propulsion between the baseline mission propulsion of 43 mm and the eliminated mission propulsion of 34 mm is calculated to be 9 mm. The propulsion difference of 9 mm is compared with the target propulsion of 42 mm in the corresponding continuous action interval to obtain the action contribution value of the torso in the corresponding continuous action interval of 21.4%. The action contribution values ​​of the right hand, forearm, upper arm and shoulder are calculated separately according to the same process, and each action contribution value is associated with body part labels and continuous action intervals. Based on the time sequence of continuous action intervals and body part identifiers, the action contribution values ​​corresponding to the target body part and each non-target body part are connected to generate a body part action contribution sequence.

[0024] In this embodiment, the generation of the action compensation event includes: The task completion sequence is obtained from the motion data of the motion task. The task completion status in the task completion sequence is matched with the continuous motion interval in the body part motion contribution sequence. The change in motion contribution of the target body part and each non-target body part between adjacent continuous motion intervals is calculated to generate the body part motion contribution change sequence. Specifically, according to the time sequence of continuous action intervals, the action contribution value of the subsequent continuous action interval is subtracted from the action contribution value of the previous continuous action interval. The action contribution value of the right hand in the adjacent continuous action interval decreases from 42.9% to 30.9%, generating a negative 12 percentage point change in action contribution. The action contribution value of the shoulder increases from 16.7% to 23.7%, generating a 7 percentage point change in action contribution. The action contribution value of the torso increases from 21.4% to 26.4%, generating a 5 percentage point change in action contribution. The change in action contribution of each body part is associated with the corresponding continuous action interval and task completion status to generate a sequence of changes in action contribution of body parts. Within the continuous action interval where the task completion state remains complete, identify continuous action intervals from the body part action contribution change sequence where the target body part action contribution change is continuously negative and at least one non-target body part action contribution change is continuously positive, connect the continuous action intervals with adjacent times, and generate candidate action compensation segments. Specifically, the changes in action contribution corresponding to continuous action intervals 6 to 10 are arranged in chronological order. The changes in action contribution for the right hand are -8%, -12%, -9%, -9%, and -5%, respectively. The changes in action contribution for the shoulder are 8%, 7%, 4%, 0%, and 0%, respectively. The changes in action contribution for the torso are 0%, 5%, 5%, 9%, and 5%, respectively. The above changes in action contribution are mapped to the task completion sequence. The continuous action intervals 6 to 10, where the task completion status is all completed, are connected to generate candidate action compensation segments with a start time of 2100 milliseconds and an end time of 3100 milliseconds. For each candidate action compensation segment, according to the time sequence of the continuous action interval, the reduction in action contribution of the target body part and the increase in action contribution of each non-target body part are obtained interval by interval. According to the order in which the action contribution of each non-target body part begins to increase, the reduction in action contribution that has not yet been received in the current continuous action interval is matched with the increase in action contribution that has not yet been allocated to each non-target body part. The smaller of the two values ​​is taken as the action contribution received by the corresponding non-target body part in the current continuous action interval. The corresponding reduction in action contribution and increase in action contribution are deducted simultaneously. The non-target body part identifier, continuous action interval and action contribution received are associated to generate action contribution received data. Specifically, within the continuous action interval 7, the decrease in the right hand's action contribution is 12 percentage points, while the increases in the shoulder and torso's action contributions are 7 and 5 percentage points, respectively. The shoulder's action contribution begins to increase earlier than that of the torso. First, the 12 percentage points that the right hand has not yet received are matched with the 7 percentage points that the shoulder has not yet been allocated, generating a shoulder action contribution of 7 percentage points. Then, the decrease in the unreceived action contribution is updated to 5 percentage points. Next, the 5 percentage points are matched with the 5 percentage points that the torso has not yet been allocated, generating a torso action contribution of 5 percentage points. The shoulder identifier, torso identifier, continuous action interval 7, and the corresponding action contribution acceptance are associated to generate action contribution acceptance data. Based on the action contribution data, non-target body parts that continuously generate action contribution data in time are identified. The target body parts, the corresponding non-target body parts, candidate action compensation segments, and action contribution data are associated to generate action compensation events. Specifically, based on the action contribution acceptance amount corresponding to the same non-target body part connected by continuous action intervals, the action contribution acceptance process of each non-target body part is formed. The action contribution acceptance amounts of the shoulder in continuous action intervals 6 to 8 are 8 percentage points, 7 percentage points, and 4 percentage points, respectively, with a cumulative action contribution acceptance amount of 19 percentage points. The action contribution acceptance amounts of the torso in continuous action intervals 7 to 10 are 5 percentage points, 5 percentage points, 9 percentage points, and 5 percentage points, respectively, with a cumulative action contribution acceptance amount of 24 percentage points. The right hand, shoulder, torso, candidate action compensation segments, and corresponding action contribution acceptance data are associated to generate the action compensation event corresponding to the right hand.

[0025] In this embodiment, the generation of action compensation load data includes: From the action compensation event, the target body part, the corresponding non-target body part, the candidate action compensation segment and the action contribution acceptance data are obtained. The non-target body parts are arranged according to the continuous action interval in which each non-target body part first generates the action contribution acceptance amount, and the action compensation time sequence is generated. Specifically, the continuous motion intervals in which the shoulder and torso first generate motion contribution are read from the motion contribution data, and the start time of the corresponding continuous motion interval is used to represent the first acceptance time. The shoulder generates motion contribution for the first time in continuous motion interval 6, and the corresponding first acceptance time is 2100 milliseconds. The torso generates motion contribution for the first time in continuous motion interval 7, and the corresponding first acceptance time is 2300 milliseconds. The shoulder and torso are arranged in order according to the first acceptance time to generate the motion compensation time sequence. Using the target body part as the starting point for migration, the first non-target body part that generates the amount of action contribution in the action compensation sequence is connected to the target body part. When the amount of action contribution of the previous non-target body part decreases, the amount of action contribution of the next non-target body part increases, and both receive the amount of action contribution reduction of the same target body part and the corresponding continuous action intervals are connected, the previous non-target body part and the next non-target body part are connected to generate an action compensation migration pair. Specifically, the shoulder and torso are obtained according to the timing of the action compensation. The action contribution of the shoulder in continuous action interval 8 and continuous action interval 9 is 4 percentage points and 0 percentage points respectively, with the action contribution decreasing by 4 percentage points. The action contribution of the torso in the corresponding continuous action interval is 5 percentage points and 9 percentage points respectively, with the action contribution increasing by 4 percentage points. The 4 percentage point decrease in the shoulder and the 4 percentage point increase in the torso both come from the decrease in action contribution of the right hand in the corresponding continuous action interval. Moreover, continuous action interval 8 and continuous action interval 9 are connected in time, connecting the shoulder and the torso to generate a action compensation transfer pair from the shoulder to the torso. The target body parts and each action compensation transfer pair are connected sequentially according to the action compensation time sequence to generate the action compensation transfer path; Specifically, taking the right hand as the starting point of migration, the right hand is connected to the shoulder, which is the first part in the action compensation sequence to receive action contribution, and then the shoulder is connected to the torso in the action compensation migration pair. The corresponding body parts are arranged in the order of the action contribution of the right hand, shoulder and torso to generate an action compensation migration path from the right hand to the shoulder and then to the torso. For each non-target body part in the motion compensation migration path, the changes in position and joint angles are accumulated during the motion contribution period and during the period when no motion contribution is generated in the same task phase, generating actual motion volume and baseline motion volume. The difference between the two is determined as additional motion volume. The moment when the first motion contribution is generated is taken as the compensation start moment, and the moment when the motion contribution reaches zero is taken as the compensation end moment. If it does not reach zero, the end moment of the motion compensation event is taken as the compensation end moment, generating the compensation duration. The number of times the same non-target body part appears in each motion compensation event is counted to generate the number of repeated participations. The additional motion volume, compensation duration, and number of repeated participations are associated with the corresponding non-target body part to generate motion compensation load data. Specifically, the cumulative positional change distance of the torso between adjacent sampling times during the motion contribution period was calculated to obtain a cumulative positional change of 140 mm. Combined with the length of the corresponding limb segment of the torso, the joint angle change at each sampling time was converted into the corresponding end position change, and the positional data was used for correction to obtain a corrected actual motion amount of 190 mm. The baseline motion amount of 80 mm during the period when the torso did not generate motion contribution during the same task phase was calculated according to the same processing procedure. The difference between the two was determined as an additional motion amount of 110 mm. The moment when the torso first generated motion contribution was 2300 milliseconds, and the end time of the motion compensation event was 3100 milliseconds. The torso still had 5 percentage points of motion contribution at the end of the motion compensation event. 3100 milliseconds was taken as the compensation end time, and a compensation duration of 800 milliseconds was generated. The torso appeared in 2 of the 3 motion compensation events corresponding to the right hand, generating 2 repeated participation counts.

[0026] In this embodiment, the generation of compensation failure response data includes: Obtain the triggering body parts, action triggering time periods, action triggering order, and task completion relationships corresponding to each interaction target in the initial somatosensory interaction task, associate them according to the interaction targets, and generate initial task structure data; Specifically, the initial motion-sensing interaction task includes interaction target A, interaction target B, and interaction target C. The triggering body part corresponding to the three interaction targets is the right hand. The action triggering time of interaction target A is 2100 ms to 2500 ms, the action triggering time of interaction target B is 2500 ms to 2900 ms, and the action triggering time of interaction target C is 2900 ms to 3300 ms. The action triggering order is interaction target A, interaction target B, and interaction target C. After interaction target A is completed, interaction target B is entered, and after interaction target B is completed, interaction target C is entered. The interaction target identifier, triggering body part, action triggering time, action triggering order, and task completion relationship are associated to generate the initial task structure data. According to the connection order from the target body part to each non-target body part in the action compensation migration path, the action contribution period corresponding to each non-target body part is matched with the action triggering period of each interactive target in the initial task structure data to determine the interactive target and action triggering period when each non-target body part participates in action compensation, and generate compensation path corresponding data. Specifically, the shoulder continuously generates motion contribution during continuous motion intervals 6 to 8, with a corresponding motion contribution period of 2100 to 2700 milliseconds. This overlaps with the motion triggering period of interactive target A for 400 milliseconds and with the motion triggering period of interactive target B for 200 milliseconds. The shoulder is associated with interactive target A based on the overlap duration. The torso continuously generates motion contribution during continuous motion intervals 7 to 10, with a corresponding motion contribution period of 2300 to 3100 milliseconds. This overlaps with the motion triggering periods of interactive target A, interactive target B, and interactive target C for 200 milliseconds, 400 milliseconds, and 200 milliseconds, respectively. The torso is associated with interactive target B based on the overlap duration. By associating the shoulder, torso, corresponding interactive target, and motion triggering period, corresponding data for the compensatory path is generated. Process the data corresponding to the compensation path sequentially along the motion compensation migration path. According to the connection order of each non-target body part participating in motion compensation, obtain the interaction target corresponding to the adjacent non-target body part. Move the action triggering time of the interaction target corresponding to the next non-target body part forward to the action triggering time of the interaction target corresponding to the previous non-target body part that has not yet been completed, so that the action triggering time of the adjacent interaction targets overlaps with each other. Adjust the corresponding interaction targets from sequential triggering to simultaneous triggering. Regenerate the action triggering order according to the adjusted action triggering time and generate task structure adjustment data. Specifically, in the motion compensation transfer path, the shoulder and torso correspond to interaction target A and interaction target B, respectively. The original motion triggering period for interaction target A is 2100 to 2500 milliseconds, and the original motion triggering period for interaction target B is 2500 to 2900 milliseconds. The motion triggering start time for interaction target B is moved forward to 2300 milliseconds, while maintaining a 400-millisecond motion triggering duration, generating a motion triggering period of 2300 to 2700 milliseconds. This ensures that interaction targets A and B are simultaneously in the triggering process within 2300 to 2500 milliseconds, thus transferring the original motion triggering period based on interaction target A and B. The task execution order formed by the sequential triggering of interactive target B is adjusted to the simultaneous triggering of two interactive targets. The time connection process of the shoulder action contribution gradually decreasing and the torso action contribution increasing is compressed. The sum of the target advancement of interactive targets A and B along the right-hand task target trajectory is 68 mm, which is lower than the maximum advancement of 76 mm completed by the right hand in the initial somatosensory interactive task. The original triggering body parts of interactive targets A and B are associated with the adjusted action triggering time period, and the action triggering order is rearranged according to the adjusted action triggering time period to generate task structure adjustment data. The task structure adjustment data replaces the triggering body part, action triggering time period and action triggering order of the corresponding interactive target in the initial task structure data, while keeping the other interactive targets and task completion relationships unchanged, to generate a compensatory failure somatosensory interactive task. Specifically, the triggering body parts, action triggering time periods, and action triggering order corresponding to interactive target A and interactive target B are read from the task structure adjustment data, and the corresponding content in the initial task structure data is replaced. The triggering body parts and the action triggering time period of 2100 milliseconds to 2500 milliseconds for interactive target A remain unchanged. Interactive target B uses the original triggering body parts, and the action triggering time period is replaced from 2500 milliseconds to 2900 milliseconds to 2300 milliseconds to 2700 milliseconds. The interactive target C, which has not undergone task structure adjustment, and its task completion relationship remain unchanged, and a compensatory failure somatosensory interactive task is generated. The system collects position data, joint angle data, interactive target trigger status, and task completion status of the target body parts and non-target body parts during the execution of the compensatory failure somatosensory interaction task. The data is then connected according to the sampling time and body part identifier to generate the compensatory failure target body part action sequence, compensatory failure non-target body part action sequence, compensatory failure interactive target trigger sequence, and compensatory failure task completion sequence. The system then associates these sequences to generate compensatory failure response data. Specifically, positional and joint angle data of the right hand, shoulder, torso, and unaffected arm are collected at 20 sampling times per second. These data are then connected according to body part identifiers and sampling times to generate a compensatory failure target body part action sequence and a compensatory failure non-target body part action sequence. The interaction target trigger states of interaction target A, interaction target B, and interaction target C at each sampling time are connected according to the interaction target identifier to generate a compensatory failure interaction target trigger sequence. The task completion states at each sampling time are connected according to the task execution time to generate a compensatory failure task completion sequence. The above four types of sequences are then associated according to the sampling times to generate compensatory failure response data.

[0027] In this embodiment, the generation of human health status monitoring and analysis results includes: According to the target body part, the motion compensation events, motion compensation migration paths and motion compensation load data are matched to obtain the number of motion compensation events corresponding to the target body part, count the number of corresponding motion compensation migration paths, and count the number of non-target body parts in each motion compensation migration path. Combine the extra exercise volume, compensation duration and repeated participation of each non-target body part to generate motion compensation feature data. Specifically, the right hand corresponds to 3 motion compensation events and 2 motion compensation transfer paths. The first motion compensation transfer path includes one non-target body part, the shoulder, and the second motion compensation transfer path includes two non-target body parts, the shoulder and the torso. The additional motion of the shoulder is 90 mm, the compensation duration is 700 ms, and the number of repetitions is 3. The additional motion of the torso is 110 mm, the compensation duration is 800 ms, and the number of repetitions is 2. The above data are associated according to the right hand identifier to generate motion compensation feature data corresponding to the right hand. From the compensation failure response data, obtain the action sequence of the target body part, the action sequence of the non-target body part, the interaction target triggering sequence, and the task completion sequence of the compensation failure. Statistically count the number of actions completed by the target body part, the number of interaction targets triggered by the target body part, and the task completion status to generate compensation failure feature data. Specifically, the compensatory failure somatosensory interaction task includes interaction target A, interaction target B, and interaction target C, and three corresponding task actions. In the completion sequence of the compensatory failure task, the task completion status corresponding to interaction target A and interaction target C is "completed", and the task completion status corresponding to interaction target B is "incomplete". In the triggering sequence of the compensatory failure interaction target, the triggering status of interaction target A and interaction target C is "triggered", and the triggering status of interaction target B is "not triggered". The number of actions completed by the right hand (2), the number of interaction targets triggered by the right hand (2), and the task incomplete status are statistically obtained. The above data are correlated to generate compensatory failure feature data. Based on action compensation feature data, identify target body parts that repeatedly participate in action compensation, form multiple action compensation migration paths, and are not target body parts that repeatedly participate in action compensation, and generate action compensation continuous state; based on compensation failure feature data, identify target body parts that have not completed the corresponding action, have not triggered the corresponding interaction target, or have not completed the task, and generate compensation failure state. Specifically, the right hand corresponds to 3 action compensation events and 2 action compensation migration paths. The shoulder appears 3 times in the 3 action compensation events, and the torso appears in 2 of the 3 action compensation events. The number of action compensation events, the number of action compensation migration paths, the number of times the shoulder participates repeatedly (3 times), and the number of times the torso participates repeatedly (2 times) are associated with the right hand to generate the action compensation continuous state. The number of actions completed by the right hand (2), the number of interactive targets triggered (2), the interactive target B not being triggered, and the task not being completed are associated to generate the compensation failure state corresponding to the right hand. The continuous state of motor compensation and the state of compensation failure are mapped according to the target body parts. The target body parts that simultaneously have both the continuous state of motor compensation and the state of compensation failure are identified as the local motor ability decline that is masked by motor compensation. The corresponding motor compensation feature data and compensation failure feature data are correlated to generate human health status monitoring and analysis results. Specifically, the continuous state and failure state of the right hand's motion compensation are mapped to obtain 3 motion compensation events, 2 motion compensation migration paths, 90 mm of additional shoulder movement, 110 mm of additional trunk movement, 2 completed right hand movements, 2 triggered interaction targets, and the task incomplete state. In the initial somatosensory interaction task, the right hand maintains task completion by contributing to the movement of the shoulder and trunk. In the somatosensory interaction task with failure of compensation, the right hand does not complete the movement corresponding to interaction target B and does not trigger interaction target B. The above data are correlated to generate human health status monitoring and analysis results.

[0028] Example 1: To verify the feasibility of this invention in practice, it was applied to the somatosensory interactive training room of a rehabilitation training center to monitor the health status of trainees during upper limb extension, lateral movement, and contact with interactive targets. The monitored subject's right-hand range of motion and movement speed decreased, but when performing routine somatosensory interactive tasks, they were still able to trigger the interactive target and complete the task by raising their shoulders and leaning forward. Existing monitoring methods typically only record whether the interactive target was triggered, whether the task was completed, and the duration of the movement. Therefore, after completing all interactive targets, the monitored subject was easily judged to have a normal motor state, failing to detect that the decline in right-hand motor ability was masked by additional shoulder and torso movements.

[0029] During application, the motion-sensing interaction task is set with interaction target A, interaction target B, and interaction target C. The monitored object is required to use its right hand to complete the actions corresponding to the three interaction targets in sequence. The motion-sensing acquisition device continuously collects position data and joint angle data of the right hand, forearm, upper arm, shoulder, and torso, while recording the trigger status and task completion status of each interaction target. This forms the raw motion-sensing interaction data. After time synchronization, coordinate unification, and body part association of the raw motion-sensing interaction data, standard motion-sensing interaction data is generated.

[0030] The right hand, designated to perform the task action, is identified as the target body part, while the forearm, upper arm, shoulder, and torso are identified as non-target body parts. Position data and joint angle data of the right hand are connected in the sampling order to generate a sequence of actions for the target body part. Position data and joint angle data of each non-target body part are then connected to generate a sequence of actions for those non-target body parts. Finally, the interactive target trigger state and task completion state are connected to generate a task completion sequence, thus forming the motion-sensing task action data.

[0031] The target body part motion sequence and the non-target body part motion sequence are aligned with the task target trajectory corresponding to the initial somatosensory interaction task. Continuous motion intervals are divided according to adjacent sampling times, and the target motion direction and target propulsion amount are calculated for each continuous motion interval. Combining the positional changes, joint angle changes, limb lengths, and body part connectivity of each body part, the target directional motion amount corresponding to each body part is generated. One continuous motion interval is selected for analysis, and the target directional motion amounts for the right hand, forearm, upper arm, shoulder, and torso are 18 mm, 4 mm, 5 mm, 7 mm, and 9 mm, respectively.

[0032] Based on the connection relationships between the torso, shoulder, upper arm, forearm, and right hand, the target direction motion quantities of each body part are mapped to the right hand distal end, generating a baseline task propulsion quantity of 43 mm. After removing the target direction motion quantity corresponding to the torso, a new culling task propulsion quantity of 34 mm is generated for the right hand distal end, with a difference of 9 mm. This propulsion difference is compared with the target task propulsion quantity to obtain the motion contribution value corresponding to the torso. The motion contribution values ​​of the right hand, forearm, upper arm, and shoulder are calculated separately using the same method. Then, the motion contribution values ​​corresponding to each body part are connected according to the sequential order of continuous motion intervals to generate a body part motion contribution sequence.

[0033] Analysis of the action contribution sequences of different body parts revealed that, when the task completion status remained incomplete, the action contribution value of the right hand decreased from 42.9% to 30.9%, a reduction of 12 percentage points; the action contribution value of the shoulder increased from 16.7% to 23.7%, an increase of 7 percentage points; and the action contribution value of the torso increased from 21.4% to 26.4%, an increase of 5 percentage points. This indicates that while the right hand's contribution to task completion decreased, the shoulder and torso increased their contribution, ensuring that the interactive target could still be triggered, and the outcome of the routine task remained largely unchanged.

[0034] The decrease in the right hand's motion contribution was further correlated with the increase in the shoulder and torso's motion contributions. Within a continuous motion interval, the decrease in the right hand's motion contribution was 12 percentage points, while the increases in the shoulder and torso's motion contributions were 7 and 5 percentage points, respectively. First, the 7 percentage point increase in the shoulder's motion contribution was correlated with the decrease in the right hand's motion contribution, resulting in a 7 percentage point shoulder motion contribution carryover. Then, the 5 percentage point increase in the torso's motion contribution was correlated with the remaining decrease in the right hand's motion contribution, resulting in a 5 percentage point torso motion contribution carryover. This generated the motion contribution carryover data.

[0035] Based on the contribution of continuous motion intervals connecting the same non-target body parts, the shoulder's cumulative contribution is 19 percentage points, and the trunk's cumulative contribution is 24 percentage points. The shoulder generates contribution earlier than the trunk, then the shoulder's contribution gradually decreases while the trunk's contribution increases accordingly, thus forming a motion compensation transfer path that connects the right hand, shoulder, and trunk in sequence. This motion compensation transfer path indicates that the reduced motion contribution of the right hand is first taken up by the shoulder, and then continues to transfer to the trunk.

[0036] The motion load of the shoulder and trunk was calculated, and the additional motion of the shoulder was 90 mm, with 3 repetitions of motion compensation. The actual motion of the trunk was 190 mm, the baseline motion was 80 mm, the additional motion was 110 mm, and 2 repetitions of motion compensation were performed. This allows for the quantification of the degree of additional motion and repetition of the shoulder and trunk during the compensation process, rather than judging motion compensation solely based on whether the body part moves.

[0037] To verify whether the right-hand local motor ability is masked by movement compensation, the movement contribution process of the shoulder and torso is mapped to interaction target A and interaction target B respectively. The task structure of the initial somatosensory interaction task is adjusted according to the movement compensation migration path. The interaction target A and interaction target B, which were originally triggered sequentially, are changed to be triggered simultaneously. This compresses the movement compensation time connection process of the shoulder participating first and the torso participating later, while keeping the interaction target C and its task completion relationship unchanged. This generates a somatosensory interaction task with compensation failure.

[0038] When the monitored object performs a compensatory haptic interaction task, the system continues to collect positional data, joint angle data, interaction target triggering status, and task completion status of the right hand, shoulder, and torso. Application logs show that the actions corresponding to interaction targets A and C can be completed and triggered, while the action corresponding to interaction target B is not completed and interaction target B is not triggered. The number of actions completed by the right hand decreased from 3 to 2, the number of triggered interaction targets decreased from 3 to 2, and the overall task completion status changed from completed to incomplete.

[0039] By analyzing comprehensive motion compensation events, motion compensation migration paths, motion compensation load data, and compensation failure response data, it was found that in the initial somatosensory interaction task, the monitored subject used the shoulder and torso to absorb the reduced motion contribution of the right hand, enabling all three interaction targets to be triggered. After compressing the time connection process of the shoulder and torso participating in motion compensation in turn, the right hand could not independently complete the action corresponding to interaction target B, thus confirming that there was a local decrease in the right hand's motor ability masked by motion compensation.

[0040] The above applications demonstrate that this invention can identify the compensatory process of decreased movement contribution from target body parts and increased movement contribution from non-target body parts while still completing a conventional somatosensory interaction task. It generates a movement compensation migration path through the movement contribution succession relationship and then verifies the decline in localized motor ability using a somatosensory interaction task where compensation fails. Compared to methods that rely solely on task completion status, this invention can more accurately detect upper limb motor ability declines masked by additional shoulder and trunk movements, improving the accuracy and relevance of human health status monitoring and analysis results.

[0041] Table 1. Comparison of the overall performance of different methods for monitoring human health status

[0042] As shown in Table 1, the task completion determination method takes 0.72 seconds per analysis, which is relatively fast. However, it only judges whether the interactive target is triggered and whether the task is completed. When the monitored object completes all interactive targets with the help of the shoulder and torso, the method is still likely to give a normal result. Therefore, the accuracy of the concealed local movement ability reduction recognition is only 70.8%, the false negative rate reaches 25.4%, and it cannot identify motion compensation events and motion compensation migration paths.

[0043] The motion trajectory deviation method can detect the deviation between the trajectory of the target body part's end point and the trajectory of the task target, improving the recognition accuracy to 78.6% and reducing the false negative rate to 17.8%. However, raising the shoulder or leaning the torso forward can propel the right hand's end point to continue approaching the interactive target, keeping the end point trajectory within an acceptable range. Therefore, the method's motion compensation event recognition rate is 54.2%, and the motion compensation migration path recognition rate is only 26.8%.

[0044] Joint angle analysis can detect abnormally involved movements by changing the angles of the shoulder, elbow, and trunk joints. Its movement compensation event recognition rate reaches 68.7%, and the consistency rate with the manual review results reaches 84.9%. However, joint angle changes can only reflect whether body parts are involved in the movement, and it is difficult to determine their actual contribution to the task progress process, and it is also difficult to determine whether the movement contribution is transferred between different body parts. Therefore, the movement compensation transfer path recognition rate is 41.5%, and the measurement error of the compensation extra movement is 27.4%.

[0045] The present invention achieves an accuracy rate of 92.3% in recognizing the decrease in concealed local movement ability, which is 21.5 percentage points higher than the task completion determination method; the false negative rate is reduced to 6.4%, which is 19.0 percentage points lower than the task completion determination method; the recognition rate of movement compensation events and the recognition rate of movement compensation migration paths reach 91.6% and 89.4% respectively, indicating that the present invention can recognize the decrease in the movement contribution of the target body part and the process of non-target body parts successively taking over the movement contribution.

[0046] The measurement error of the compensatory extra motion volume of the present invention is 8.7%, which is lower than the 38.6% of the motion trajectory deviation method and the 27.4% of the joint angle analysis method. The reason is that the present invention maps the position changes and joint angle changes of body parts to the target direction of the task, and calculates the motion contribution by removing the motion volume corresponding to the target direction of the body parts to be analyzed one by one, thereby distinguishing between general coordinated movements and compensatory movements that actually promote the completion of the task.

[0047] Simultaneously, this invention correlates the reduction in action contribution from the target body part with the increase in action contribution from non-target body parts to generate action contribution transfer data. It then combines this with the order in which different body parts begin to receive action contributions to form a motion compensation migration path. Furthermore, by adjusting the triggering time and order of the interactive target, a motion-sensory interaction task with compensation failure is generated to verify the decline in localized motor ability masked by motion compensation. Therefore, the consistency rate with the results of manual review reaches 92.7%.

[0048] The average analysis time per task in this invention is 1.08 seconds. Although this is 0.36 seconds longer than the task completion determination method, it remains within the second range. The increased processing time is mainly used for motion contribution calculation, motion compensation migration analysis, and compensation failure response analysis. Compared to improvements in recognition accuracy, false negative rate, migration path recognition rate, and measurement error, this computational overhead is relatively small and can meet the need for quickly generating human health status monitoring and analysis results after the completion of a motion-sensing interaction task.

[0049] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for monitoring and analyzing human health status based on somatosensory interaction, characterized in that, Includes the following steps: Collect raw somatosensory interaction data during the initial somatosensory interaction task performed by the monitored object, synchronize the somatosensory interaction data in time, unify the coordinates and associate the raw somatosensory interaction data with body parts, and generate standard somatosensory interaction data; Based on standard motion-sensing interaction data, human motion processes are divided into target body parts and non-target body parts to generate motion-sensing task motion data. Based on the motion data of the motion-sensing task and the target trajectory corresponding to the initial motion-sensing interaction task, the motion contribution of the target body parts and non-target body parts to the task completion process is calculated respectively, and a sequence of body part motion contribution is generated. Based on the body part motion contribution sequence and somatosensory task motion data, identify the motion process in which the motion contribution of the target body part decreases, the motion contribution of the non-target body part increases, and the task completion status remains complete, and generate motion compensation events. Based on the motion compensation events, motion compensation migration paths are generated, and the additional motion, compensation duration, and number of repetitions for each non-target body part are calculated to generate motion compensation load data. Based on the motion compensation migration path, the task structure of the initial somatosensory interaction task is changed to generate a somatosensory interaction task with compensation failure, and the motion response during the collection period is used to generate compensation failure response data. Based on motion compensation events, motion compensation migration paths, motion compensation load data, and compensation failure response data, the system identifies localized declines in motor ability masked by motion compensation, and generates human health status monitoring and analysis results.

2. The method for monitoring and analyzing human health status based on somatosensory interaction according to claim 1, characterized in that, The raw data for the somatosensory interaction includes human motion data and task execution data; The raw motion-sensing interaction data is synchronized in time, unified in coordinates, and associated with body parts to generate standard motion-sensing interaction data.

3. The method for monitoring and analyzing human health status based on somatosensory interaction according to claim 1, characterized in that, The generation of the motion-sensing task action data includes: Obtain body part identifiers, position data, joint angle data, and task execution status corresponding to each sampling moment from standard somatosensory interaction data; In the initial somatosensory interaction task, the body part that is designated to perform the task action is identified as the target body part, and the body parts that undergo positional or joint angle changes during human movement, other than the target body parts, are identified as non-target body parts. The target body part's motion sequence is generated by connecting the position data and joint angle data corresponding to the target body part according to the sampling time, and the non-target body part's motion sequence is generated by connecting the position data and joint angle data corresponding to each non-target body part according to the body part identifier. The task execution status is connected according to the sampling time to generate a task completion sequence. The target body part action sequence, the non-target body part action sequence and the task completion sequence are associated according to the sampling time to generate motion-sensing task action data.

4. The method for monitoring and analyzing human health status based on somatosensory interaction according to claim 1, characterized in that, The generation of the body part movement contribution sequence includes: Align the sequence of actions of the target body parts and the sequence of actions of non-target body parts with the trajectory of the mission target in time, divide the continuous action intervals, and determine the target movement direction and target propulsion in each continuous action interval; The changes in position and joint angles of each body part within each continuous movement interval are obtained. The changes in position are projected onto the target movement direction, and the changes in joint angles are combined to generate the movement amount in the target direction. Based on the connection relationship between body parts, the motion quantities in each target direction are transferred to the target body part to generate the baseline task progress quantity. The motion quantities in each target direction corresponding to each body part are removed in turn, and the remaining motion quantities in the target direction are used to generate the culling task progress quantity. The difference between the baseline task advance and the elimination task advance is calculated. The difference is compared with the target advance to generate the action contribution value corresponding to each body part. The action contribution values ​​are then connected in the time sequence of continuous action intervals to generate the action contribution sequence of body parts.

5. The method for monitoring and analyzing human health status based on somatosensory interaction according to claim 1, characterized in that, The generation of the action compensation event includes: The task completion status in the task completion sequence is mapped to the continuous action interval in the body part action contribution sequence. The change in action contribution of each body part between adjacent continuous action intervals is calculated to generate a body part action contribution change sequence. Within a continuous action interval where the task completion state remains complete, identify a continuous action interval where the action contribution of the target body part continuously decreases and the action contribution of at least one non-target body part continuously increases, and connect the continuous action intervals that are adjacent in time to generate candidate action compensation segments. According to the time sequence of continuous action intervals, the decrease in the action contribution of the target body part is matched with the increase in the action contribution of each non-target body part in turn to generate the action contribution acceptance amount of each non-target body part, and the non-target body part identifier and continuous action interval are associated to generate action contribution acceptance data. Based on the motion contribution reception data, non-target body parts that continuously generate motion contribution reception volume are identified. The target body parts, non-target body parts, candidate motion compensation segments, and motion contribution reception data are associated to generate motion compensation events.

6. The method for monitoring and analyzing human health status based on somatosensory interaction according to claim 1, characterized in that, The generation of the action compensation load data includes: Data on target body parts, non-target body parts, and action contribution reception are obtained from action compensation events. Action compensation time sequence is generated according to the order in which each non-target body part first generates action contribution reception. Using the target body part as the starting point of migration, the first non-target body part is connected to the target body part. When the motion contribution of the previous non-target body part decreases and the motion contribution of the next non-target body part increases, and the corresponding continuous motion intervals of the two are connected, the previous non-target body part and the next non-target body part are connected to generate a motion compensation migration pair. Connect the target body parts and each action compensation transfer pair according to the action compensation sequence to generate the action compensation transfer path; For each non-target body part in the motion compensation migration path, the actual amount of motion during the motion contribution period and the baseline amount of motion within the same task phase are calculated. The difference between the two is determined as the additional amount of motion. Combining the time of generation of the motion contribution, the time of zeroing, and the number of occurrences in each motion compensation event, the compensation duration and the number of repeated participations are generated. The additional amount of motion, the compensation duration, and the number of repeated participations are associated with the corresponding non-target body parts to generate motion compensation load data.

7. The method for monitoring and analyzing human health status based on somatosensory interaction according to claim 1, characterized in that, The generation of the compensation failure response data includes: Obtain the triggering body parts, action triggering time periods, action triggering order, and task completion relationships corresponding to each interaction target in the initial somatosensory interaction task, associate them according to the interaction targets, and generate initial task structure data; According to the connection order from the target body part to each non-target body part in the action compensation migration path, the action contribution period corresponding to each non-target body part is matched with the action triggering period of each interactive target in the initial task structure data to determine the interactive target and action triggering period when each non-target body part participates in action compensation, and generate compensation path corresponding data. Process the data corresponding to the compensation path sequentially along the motion compensation migration path. According to the connection order of each non-target body part participating in motion compensation, obtain the interaction target corresponding to the adjacent non-target body part. Move the action triggering time of the interaction target corresponding to the next non-target body part forward to the action triggering time of the interaction target corresponding to the previous non-target body part that has not yet been completed, so that the action triggering time of the adjacent interaction targets overlaps with each other. Adjust the corresponding interaction targets from sequential triggering to simultaneous triggering. Regenerate the action triggering order according to the adjusted action triggering time and generate task structure adjustment data. The task structure adjustment data replaces the triggering body part, action triggering time period and action triggering order of the corresponding interactive target in the initial task structure data, while keeping the other interactive targets and task completion relationships unchanged, to generate a compensatory failure somatosensory interactive task. The system collects positional data, joint angle data, interactive target trigger status, and task completion status of the target body parts and non-target body parts during the execution of the compensatory failure somatosensory interaction task. These data are then connected according to the sampling time and body part identifier to generate the compensatory failure target body part action sequence, compensatory failure non-target body part action sequence, compensatory failure interactive target trigger sequence, and compensatory failure task completion sequence. These sequences are then correlated to generate compensatory failure response data.

8. The method for monitoring and analyzing human health status based on somatosensory interaction according to claim 1, characterized in that, The generation of the human health status monitoring and analysis results includes: According to the target body part, the motion compensation events, motion compensation migration paths and motion compensation load data are matched, and the number of motion compensation events, the number of motion compensation migration paths, the number of non-target body parts, the extra movement volume of each non-target body part, the compensation duration and the number of repeated participations are counted to generate motion compensation feature data. The action sequences, interaction target trigger sequences, and task completion sequences are obtained from the compensation failure response data. The number of actions completed by the target body parts, the number of interaction targets triggered, and the task completion status are counted to generate compensation failure feature data. The continuous state of action compensation is identified based on action compensation feature data, and the failure state of compensation is identified based on compensation failure feature data. Based on the continuous state and failure state of the corresponding movement compensation for the target body part, the target body part that simultaneously has both states is identified as having a localized decline in motor ability that is masked by movement compensation. The corresponding movement compensation feature data and compensation failure feature data are then associated to generate human health status monitoring and analysis results.

Images