A virtual reality-based health management monitoring method and system

Abstract

The application provides a kind of health management monitoring method and system based on virtual reality, it is related to health state analysis technical field.The method comprises: obtaining the behavior record data and physiological record data of a plurality of users in the process of virtual reality scene experience, carrying out behavior mutation detection and physiological mutation detection to each user, determine a plurality of behavior mutation points and physiological mutation points and fuse into a plurality of joint change points, extract the plurality of joint state sequences of user and carry out joint state pattern recognition, generate a plurality of joint state patterns and determine a plurality of transition patterns;Construct the state transition fingerprint of each transition pattern, after collecting the target behavior data and target physiological data of the user to be monitored, based on a plurality of state transition fingerprints, the health management monitoring of the user to be monitored is carried out, and the health monitoring result of the user to be monitored is obtained.The application realizes a kind of continuous health management monitoring based on the multi-modal user state change of virtual reality scene.

Description

Technical Field

[0001] This invention relates to the field of health status analysis technology, and in particular to a health management and monitoring method and system based on virtual reality. Background Technology

[0002] With the continuous development of virtual reality technology, virtual reality devices are gradually being applied to various fields such as entertainment, training, interactive experiences, and health management. When users wear virtual reality devices, they can complete various interactive operations or experience content in an immersive virtual scene.

[0003] During the use of virtual reality devices, user behavior and physiological data can be acquired synchronously, such as changes in user movements, posture characteristics, and physiological parameters. This data is generated naturally during the user's normal experience with virtual reality content and can reflect changes in the user's state in different experience scenarios, providing a potential data foundation for health status analysis.

[0004] Existing health management technologies based on virtual reality (VR) focus primarily on analyzing certain behavioral characteristics or physiological indicators of users during the experience. However, data generated during VR experiences is multi-dimensional, continuous, and dynamically changing. Different types of data exhibit correlations, and different users show significant differences in behavior and physiological characteristics. Continuous state analysis and cross-user comparisons of this complex data could provide a more stable and objective portrayal of changes in a user's health status during the experience. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention proposes a health management monitoring method and system based on virtual reality. This method and system are used to uniformly process and analyze the multimodal behavioral and physiological data generated during the user's experience using virtual reality devices, thereby enabling continuous monitoring of changes in the user's health status.

[0006] The first aspect of this invention provides a health management and monitoring method based on virtual reality, comprising: Acquire behavioral and physiological data of multiple users during virtual reality scene experience, construct multiple behavioral and physiological change sequences for each user, perform behavioral and physiological mutation detection for each user, and identify multiple behavioral and physiological mutation points for each user. Multiple behavioral and physiological mutation points of users are fused to determine multiple joint change points, and multiple joint state sequences of users are extracted from multiple behavioral and physiological change sequences based on multiple joint change points; Perform joint state pattern recognition on multiple joint state sequences to generate multiple joint state patterns for multiple users. Extract the joint state transition path for each user based on the joint state pattern and determine multiple transition patterns based on the multiple joint state transition paths. A state migration fingerprint is constructed for each migration pattern based on multiple behavioral and physiological change sequences. After collecting the target behavioral and physiological data of the user to be monitored, the user's health management is monitored based on the multiple state migration fingerprints to obtain the health monitoring results of the user during the virtual reality scene experience.

[0007] Preferably, determining multiple behavioral and physiological mutation points for each user includes: Construct a behavior change matrix and a physiological change matrix for each user, and perform sliding window distribution difference analysis on the behavior change matrix and the physiological change matrix respectively. Calculate the sliding behavior difference of the behavior change matrix between any two adjacent sliding windows, and calculate the sliding physiological difference of the physiological change matrix between any two adjacent sliding windows. Based on the sliding behavior difference and the sliding physiological difference, determine multiple behavioral change points and physiological change points for the user.

[0008] Preferably, fusing multiple behavioral and physiological mutation points of the user to determine multiple joint change points, and extracting multiple joint state sequences of the user from multiple behavioral and physiological change sequences based on multiple joint change points includes: The system fuses multiple behavioral and physiological mutation points for each user, including constructing a joint mutation curve for the user, performing a sliding scan on multiple mutation points in the joint mutation curve, and if the time difference between the current mutation point and the next mutation point is less than a preset tolerance threshold, then the current mutation point and the next mutation point are fused, including replacing the current mutation point and the next mutation point with the mean of the two mutation points, and outputting multiple joint change points after traversing multiple mutation points in the joint mutation curve. Multiple segment intervals of the user are constructed based on multiple joint change points. The user's behavior change matrix and physiological change matrix are extracted from multiple interval behavioral features in each segment interval, and the user's physiological change matrix and physiological features in each segment interval are extracted from multiple interval behavioral features and physiological features. A joint state sequence for each segment interval is constructed based on the multiple interval behavioral features and physiological features.

[0009] Preferably, the joint state migration path for each user is extracted based on the joint state pattern, and multiple migration patterns are determined based on multiple joint state migration paths, including: Determine the joint state pattern of the user in each segment interval, and construct the joint state transition path of the user based on the joint state pattern of the user in multiple segment intervals; Extract multiple transition patterns from each joint state transition path, where each transition pattern includes two adjacent joint state patterns in the joint state transition path; For multiple joint state patterns, multiple joint state sequences are clustered to generate multiple joint state patterns.

[0010] Preferably, constructing the state transition fingerprint for each migration pattern based on multiple behavioral change sequences and physiological change sequences includes: Determine multiple joint state migration paths to which each migration pattern belongs, and construct a migration sample library for each migration pattern. The migration sample library includes local behavior matrices and local physiological matrices extracted from multiple users containing the corresponding migration patterns. Construct the migration start state vector and migration end state vector for each user in the migration sample library with respect to the migration mode. Generate the migration progress curve for each user based on the migration start state vector and migration end state vector. This includes calculating the state migration distance based on the migration start state vector and migration end state vector, determining the user's joint migration state vector at any time, determining the user's displacement vector under the joint migration state vector based on the migration start state vector, constructing the user's migration direction vector with respect to the migration mode based on the migration start state vector and migration end state vector, calculating the projection of the user's displacement vector onto the migration direction vector, obtaining the user's migration progress value under the joint migration state vector, and constructing the user's migration progress curve based on the user's migration progress values ​​at multiple time points. The migration start feature, migration end feature, and migration path length feature of the user are extracted from the migration progress curve to generate the individual migration fingerprint of the user. The individual migration fingerprints of multiple users in the migration sample library are then merged to obtain the state migration fingerprint of the migration pattern.

[0011] Preferably, health management monitoring of the user to be monitored based on multiple state transition fingerprints includes: Based on the target behavior data and target physiological data of the user to be monitored, behavioral mutation detection and physiological mutation detection are performed on the user to be monitored to determine multiple target joint change points of the user to be monitored, and multiple target joint state sequences of multiple users to be monitored are extracted based on the target joint change points. Multiple target joint state sequences are matched with multiple joint state patterns respectively. Based on the state pattern matching results, the target migration path of the user to be monitored is determined, and the target migration pattern to which the target migration path belongs is identified. Construct a target migration fingerprint for the user to be monitored along the target migration path. Based on the state migration fingerprint of the target migration pattern, perform deviation detection on the target migration fingerprint, determine the deviation parameters of the target migration fingerprint, and generate health monitoring results for the user to be monitored regarding target behavioral data and target physiological data.

[0012] A second aspect of the present invention provides a virtual reality-based health management and monitoring system for implementing the aforementioned virtual reality-based health management and monitoring method, comprising: The sample data acquisition and analysis module is used to acquire behavioral and physiological data of multiple users during the virtual reality scene experience, construct multiple behavioral and physiological change sequences for each user, perform behavioral and physiological mutation detection for each user, and determine multiple behavioral and physiological mutation points for each user. The joint state analysis module is used to fuse multiple behavioral and physiological mutation points of a user to determine multiple joint change points, and extract multiple joint state sequences of the user from multiple behavioral and physiological change sequences based on multiple joint change points; The migration pattern recognition module is used to perform joint state pattern recognition on multiple joint state sequences, generate multiple joint state patterns corresponding to multiple users, extract the joint state migration path of each user based on the joint state pattern, and determine multiple migration patterns based on the multiple joint state migration paths. The health monitoring and analysis module is used to construct a state migration fingerprint for each migration pattern based on multiple behavioral change sequences and physiological change sequences. After collecting the target behavioral data and target physiological data of the user to be monitored, the module performs health management monitoring on the user to be monitored based on multiple state migration fingerprints, and obtains the health monitoring results of the user to be monitored during the virtual reality scene experience.

[0013] The present invention has the following beneficial effects: This invention jointly processes multimodal behavioral and physiological data generated during virtual reality scene experiences, integrates the temporal characteristics of behavioral and physiological changes, constructs a joint state sequence, and forms representative state transition patterns and their migration fingerprints based on group data. This enables continuous characterization and comparative analysis of user state change processes. Through deviation representation based on migration fingerprints and deviation detection mechanisms based on migration fingerprints, subtle changes in the user state transition process are quantitatively described, providing an objective representation of state evolution deviations. This helps support continuous monitoring and analysis of user state changes in health management-related scenarios. It comprehensively analyzes the dynamic characteristics of the user experience in virtual scene processes from the perspective of state change processes, providing an objective monitoring method based on state change processes for health management-related applications. Attached Figure Description

[0014] Figure 1 This is a flowchart illustrating a health management and monitoring method based on virtual reality provided in one embodiment of the present invention.

[0015] Figure 2This is a schematic diagram of a health management and monitoring system based on virtual reality provided in one embodiment of the present invention. Detailed Implementation

[0016] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for explaining the invention and are not intended to limit the invention.

[0017] Please see Figure 1 This document illustrates a flowchart of a virtual reality-based health management and monitoring method according to an embodiment of the present invention. This method can be applied to virtual reality device usage scenarios, and by processing and analyzing data naturally generated by users during their virtual reality experience, it enables the monitoring and management of users' health status. The method includes the following steps: Step S1: Obtain behavioral and physiological data of multiple users during their virtual reality experience, construct multiple behavioral and physiological change sequences for each user, perform behavioral and physiological mutation detection on each user, and determine multiple behavioral and physiological mutation points for each user.

[0018] Specifically, during the use of virtual reality devices, behavioral and physiological data generated by the user while experiencing virtual reality scenes can be collected. Behavioral data may include, but is not limited to, information such as changes in the user's movements, postures, and interactive operation trajectories. For example, this includes head posture data such as the user's head position, orientation, or posture changes in three-dimensional space, and hand posture data such as the user's hand position, posture, movement trajectory, or range of motion. Physiological data may include, but is not limited to, continuous changes related to the user's physiological state, such as heart rate, blood pressure, respiratory rate, skin conductance, and body surface temperature.

[0019] After acquiring behavioral and physiological data from multiple users, the data were preprocessed, including clearing and alignment, and multiple behavioral and physiological change sequences were constructed for each user in chronological order. Based on this, behavioral and physiological mutation detection was performed on each user to identify the locations of significant behavioral and physiological changes in continuous data, thus obtaining multiple behavioral and physiological mutation points for each user within the collected data.

[0020] Optionally, to determine behavioral and physiological mutation points, multiple behavioral and physiological change sequences for each user are first combined to construct behavioral and physiological change matrices for each user. Then, a sliding window distribution difference analysis is performed on the behavioral and physiological change matrices to calculate the sliding behavioral differences between any two adjacent sliding windows, and the sliding physiological differences between any two adjacent sliding windows.

[0021] In this process, the distance between the local matrices under any two adjacent sliding windows in the behavior change matrix is ​​calculated, for example, represented by Euclidean distance. This yields the difference in sliding behavior between any two adjacent sliding windows. If the difference in sliding behavior is greater than a preset behavior difference threshold, it indicates that the user's behavior state has changed for these two sliding windows, and the boundary point between the two windows is recorded as one of the behavior abrupt change points. Similarly, for the user's physiological change matrix, the physiological difference in sliding between any two adjacent sliding windows is calculated in the same way. If the physiological difference in sliding behavior is greater than a preset physiological difference threshold, the boundary point between the two windows is recorded as one of the physiological abrupt change points. Through the above method, multiple behavior abrupt change points and physiological abrupt change points for each user are obtained.

[0022] Step S2: Fuse multiple behavioral and physiological mutation points of the user to determine multiple joint change points, and extract multiple joint state sequences of the user from multiple behavioral and physiological change sequences based on multiple joint change points.

[0023] Specifically, multiple behavioral mutation points and multiple physiological mutation points are fused together over time. By comprehensively considering the consistency or correlation between behavioral and physiological changes over time, multiple joint change points are determined to characterize the temporal boundary of the user's overall state during the virtual reality experience.

[0024] After identifying multiple joint change points, the multiple behavioral change sequences and physiological change sequences are segmented according to the time interval formed by adjacent joint change points, thereby extracting multiple joint state sequences for each user. Each joint state sequence corresponds to a relatively continuous and stable state stage of the user during the virtual reality experience, which can reflect the comprehensive performance of the user's behavioral and physiological characteristics during that stage.

[0025] Optionally, the fusion of behavioral and physiological mutation points includes merging multiple behavioral and physiological mutation points in chronological order to construct a joint mutation curve for the user, thereby describing multiple mutation locations corresponding to changes in user behavior and physiology on a unified time axis. A sliding scan process is then performed on multiple mutation points in the joint mutation curve. Specifically, the current mutation point in the joint mutation curve is sequentially selected along with its temporally adjacent next mutation point, and the time difference between them is calculated.

[0026] When the time difference between the current mutation point and the next mutation point is less than a preset tolerance threshold, the two mutation points are considered to have high temporal consistency, indicating that the user's behavioral and physiological changes occurred basically synchronously within that time period. This involves fusing the current and next mutation points. The average time of the current and next mutation points can be used to replace them, generating a new fused mutation point to represent the location where the user's overall state changed within that time period. By sequentially performing the above sliding scan and fusion process on multiple mutation points in the joint mutation curve, after traversing all mutation points in the joint mutation curve, the fused joint change points are output.

[0027] To construct the joint state sequence, based on the aforementioned multiple joint change points, adjacent joint change points are used as boundaries in chronological order to construct multiple segment intervals for each user during the virtual reality experience. Each segment interval corresponds to a continuous time interval between two adjacent joint change points, representing a relatively stable phase in the user's overall state within that time interval.

[0028] After identifying multiple user segment intervals, corresponding interval features are extracted from the user's behavior change matrix and physiological change matrix for each segment interval. Specifically, for each segment interval, multiple interval behavioral features reflecting user behavior characteristics are extracted from the behavior change matrix, and multiple interval physiological features reflecting user physiological state characteristics are extracted from the physiological change matrix. These are used to characterize the user's overall behavioral performance and physiological changes within that segment interval. The interval behavioral and physiological features can employ common statistical features representing stability, dynamic intensity, etc., such as the mean, standard deviation, RMS (root mean square), and jerk (acceleration) parameters corresponding to multiple behavioral and physiological change sequences within a specific segment interval. The extracted interval behavioral and physiological features are then combined to construct the joint state sequence corresponding to that segment interval.

[0029] Through the above steps, based on the temporal consistency of behavioral and physiological changes, the user's multimodal data can be fused to form multiple joint state sequences that can reflect the overall state change process of the user, thus obtaining the joint state sequence corresponding to each user in multiple segment intervals.

[0030] Step S3: Perform joint state pattern recognition on multiple joint state sequences to generate multiple joint state patterns corresponding to multiple users. Extract the joint state migration path for each user based on the joint state pattern, and determine multiple migration patterns based on the multiple joint state migration paths.

[0031] Specifically, joint state sequences from multiple users are analyzed to identify those with similar characteristics and categorize them into the same joint state pattern. Different joint state patterns represent the typical state types exhibited by a group of users in terms of behavioral and physiological characteristics during a virtual reality experience.

[0032] After obtaining the joint state pattern for each user, a joint state transition path is constructed based on the user's chronological order during the virtual reality experience. This path describes the user's evolution between different joint state patterns. Furthermore, the joint state transition paths of multiple users are summarized and analyzed to identify common state transition processes, thereby determining multiple transition patterns.

[0033] For identifying joint state patterns, clustering can be performed on joint state sequences from multiple users. Methods such as k-means, DBSCAN, or hierarchical clustering can be used to group joint state sequences with similar characteristics into the same class, thus forming multiple joint state patterns. Each joint state pattern corresponds to a class of states that exhibit similar behaviors and physiological characteristics.

[0034] For the extraction of joint state transition paths and the generation of transition patterns, after clustering to generate multiple joint state patterns, for each user, the joint state pattern corresponding to that user in each segment interval is determined. Then, according to the time sequence of the user's virtual reality experience, i.e., the temporal order of the joint state patterns corresponding to multiple segment intervals, these joint state patterns are arranged sequentially to construct the user's joint state transition path. Further, multiple transition patterns contained within these joint state transition paths are extracted.

[0035] For any joint state transition path, the state transition relationship formed by two adjacent joint state patterns in the path is extracted, and this state transition relationship is taken as a transition pattern. This is used to characterize the basic transition form of a user evolving from one joint state pattern to another during a virtual reality experience. Furthermore, the transition patterns extracted from joint state transition paths of multiple users can be summarized and analyzed to identify recurring transition patterns as representative transition patterns at the group level. Transition patterns that appear with a frequency reaching a preset threshold in the user group are then selected, and noise information appearing only in a few individuals is removed. Finally, multiple transition patterns extracted from multiple joint state transition paths are obtained.

[0036] Step S4: Construct state transition fingerprints for each transition mode based on multiple behavioral change sequences and physiological change sequences. After collecting target behavioral data and target physiological data of the user to be monitored, conduct health management monitoring of the user to be monitored based on multiple state transition fingerprints to obtain the health monitoring results of the user to be monitored during the virtual reality scene experience.

[0037] Specifically, for each migration pattern, a state migration fingerprint is constructed by integrating multiple user behavior change sequences and physiological change sequences to characterize the features of that migration pattern. This fingerprint is used to describe the overall characteristics of a user's behavior and physiological changes during a specific state migration process.

[0038] In practical applications, when target behavior and physiological data of the user to be monitored during their virtual reality experience are collected, the collected target data is processed according to the steps described above for processing data from multiple users, constructing a target migration fingerprint for the user to be monitored. This target migration fingerprint is then compared and analyzed with multiple pre-constructed state migration fingerprints to monitor the user's health status during their virtual reality experience, yielding corresponding health monitoring results.

[0039] Optionally, for constructing the state transition fingerprint of a transition mode, first determine the multiple joint state transition paths to which each transition mode belongs, and then construct a transition sample library for each transition mode.

[0040] The migration sample library includes local behavior matrices and local physiological matrices extracted from multiple users containing the corresponding migration patterns. In this process, multiple joint state migration paths containing the migration pattern are first determined. Based on the users corresponding to these paths, local data fragments between segments corresponding to the migration pattern are extracted from each user's behavior change matrix and physiological change matrix, resulting in each user's local behavior matrix and local physiological matrix for that migration pattern.

[0041] For any transition pattern, including the evolution from one joint state pattern as the starting state to another joint state pattern as the ending state, a user's understanding of that transition pattern can be constructed based on the user's local behavior matrix and local physiological matrix. The migration start-state vector and migration end-state vector of the migration mode. The migration start-state vector and migration end-state vector can be generated by calculating the mean vector of the local behavior matrix and the local physiological matrix, and are used to represent the user's initial state and target state in this migration mode, respectively.

[0042] Based on the initial and final state vectors of the migration, the user's state migration distance at different times is analyzed to construct the user's migration progress curve under the migration mode. In this process, the state migration distance is first calculated based on the initial and final state vectors, which can be represented by the Euclidean distance between them, representing the magnitude of the overall state change during the user's transition from the initial state to the target state. For calculating the user's state migration distance at any given time, the user's joint migration state vector at that time is first determined. Using a fixed window centered on the current time, the mean vectors of the user's local behavior matrix and local physiological matrix within that window are calculated and concatenated to obtain the user's joint migration state vector at the current time.

[0043] The user's displacement vector under the joint migration state vector is determined based on the migration start state vector. The joint migration state vector represents the user's current state, and the migration start state vector represents the initial state of the migration process. The displacement vector representing the current state is obtained by subtracting the migration start state vector representing the initial state term by term, thus representing the displacement information of the current state relative to the initial state. Simultaneously, a migration direction vector representing the user's migration mode is constructed based on the migration start state vector and the migration end state vector. This is achieved by subtracting the migration end state vector representing the target state term by term, thus representing the migration direction vector representing the displacement information of the target state relative to the initial state during the migration process.

[0044] Based on the above, the projection of the user's displacement vector onto the migration direction vector is calculated to obtain the user's migration progress value under this joint migration state vector. This value is used to quantitatively characterize the user's relative completion degree between the migration start state and the migration end state. By arranging the user's migration progress values ​​at multiple time points, a migration progress curve for the user with respect to this migration mode is constructed. The user's migration start feature, migration end feature, and migration path length feature can be extracted from the migration progress curve. The migration start feature represents the relative position or temporal sequence of the user entering the migration state, and the migration end feature represents the relative position or temporal sequence of the user completing the migration state.

[0045] For example, if the total migration time for a user in this migration mode is T, the time sequence positions corresponding to the user's first arrival at the migration start point and migration end point are determined according to the pre-set migration start point and migration end point. The migration start point and migration end point specifically represent the time when the user's state begins to change and the time when the user's state ends to change. For example, a state with a migration progress value of 0.1 is recorded as the migration start point, and a state with a migration progress value of 0.9 is recorded as the migration end point. If, within the migration time interval T, the user's migration progress value first reaches the migration start point at time t1 and first reaches the migration end point at time t2, then times t1 and t2 correspond to the user's migration start characteristic and migration end characteristic, respectively. The time difference between the two times corresponds to the user's migration path length characteristic, representing the total time used by the user to complete the migration process. The migration progress curve contains the user's migration progress values ​​at different times. Those skilled in the art can reasonably set the migration progress values ​​corresponding to the migration start point and migration end point according to actual needs.

[0046] The extracted migration initiation features, migration end features, and migration path length features are used as individual migration fingerprints for each user. Further fusion of individual migration fingerprints from multiple users in the migration sample library yields state migration fingerprints for different migration patterns. For example, for each dimension of features, corresponding quantile intervals are constructed based on the individual migration fingerprints of multiple users. Preset lower and upper quantiles are then used to construct reference quantile intervals for each dimension of features, serving as normal templates for different dimensions. This method allows for the construction of representative state migration fingerprints for each migration pattern based on the state change process of multiple users under the same migration pattern, providing a basis for subsequent health management monitoring.

[0047] Optionally, for health management monitoring of users based on multiple state transition fingerprints, firstly, based on the target behavioral data and target physiological data of the users to be monitored, behavioral mutation detection and physiological mutation detection are performed to determine multiple joint change points of the users to be monitored. Similar to steps S1 and S2 above, by real-time monitoring of the behavioral and physiological data of the users to be monitored, significant changes during the experience are detected, and multiple mutation points corresponding to the behavioral and physiological data are determined. Then, similar change points are merged in chronological order to obtain the joint change points representing the comprehensive state changes of the users to be monitored. Each joint change point corresponds to the boundary of the state changes of the users to be monitored during the virtual reality experience. Based on multiple joint change points, multiple segment intervals in the target behavioral and physiological data of the users to be monitored can be determined. Similar to the extraction method of joint state sequences, multiple joint state sequences of multiple users to be monitored can be extracted based on the joint change points.

[0048] Multiple target joint state sequences are matched with multiple joint state patterns obtained from the aforementioned clustering. During the matching process, the similarity between the target joint state sequence and each joint state pattern can be calculated. Specifically, the average value of the joint state sequences of multiple users in the joint state pattern can be taken to represent a specific joint state pattern. Based on the similarity, the joint state pattern to which the user to be monitored belongs in each segment interval is determined, thereby generating the target migration path of the user to be monitored based on the sorting of multiple segment intervals.

[0049] Extract any two adjacent joint state patterns from the target migration path to obtain one or more target migration patterns corresponding to the target migration path. Similar to the construction of individual migration fingerprints, one or more target migration fingerprints of the user to be monitored under the target migration path can be constructed based on target behavioral data and target physiological data. Deviation detection is then performed on the target migration fingerprints based on the state migration fingerprints of the target migration patterns. Specifically, the state migration fingerprint corresponding to the target migration fingerprint among multiple state migration fingerprints is determined, and the degree of deviation of the target migration fingerprint from its corresponding state migration fingerprint on different feature terms is calculated. For example, for the migration path length, the migration path length feature of the user to be monitored is determined, and its local deviation index relative to the reference quantile interval of the migration path length in the state migration fingerprint is calculated. This includes calculating the absolute value of the distance between the migration path length feature and the center point of the reference quantile interval, recording the ratio between the absolute value of the distance and the length of the reference quantile interval as the local deviation index of the migration path length, and calculating the mean of the local deviation indices of the target migration fingerprint on multiple feature terms as the deviation parameter of the target migration fingerprint.

[0050] Finally, health monitoring results for the monitored user regarding target behavioral and physiological data are generated based on the deviation parameters. The degree of deviation between the target migration fingerprint and the state migration fingerprint is used to analyze whether the user's health status meets expectations. If the monitored user deviates significantly during the migration process, it indicates a potential health abnormality. The deviation parameters characterize the degree of matching between the monitored user's state migration process during the virtual reality scene experience and the migration pattern constructed based on a large number of normal user samples. This provides an objective representation of the monitored user's state changes during the virtual reality scene experience, offering data support for subsequent health management analysis or related applications. Health monitoring results including deviation parameters enable continuous recording and comparative analysis of user state changes during the virtual reality scene experience. This allows relevant health management personnel to provide corresponding health management or adjustment suggestions based on the differences in the degree of deviation among different users, supporting long-term health management applications.

[0051] Please see Figure 2 The diagram illustrates a structural schematic of a virtual reality-based health management and monitoring system according to an embodiment of the present invention. The system includes: The sample data acquisition and analysis module is used to acquire behavioral and physiological data of multiple users during the virtual reality scene experience, construct multiple behavioral and physiological change sequences for each user, perform behavioral and physiological mutation detection for each user, and determine multiple behavioral and physiological mutation points for each user. The joint state analysis module is used to fuse multiple behavioral and physiological mutation points of a user to determine multiple joint change points, and extract multiple joint state sequences of the user from multiple behavioral and physiological change sequences based on multiple joint change points; The migration pattern recognition module is used to perform joint state pattern recognition on multiple joint state sequences, generate multiple joint state patterns corresponding to multiple users, extract the joint state migration path of each user based on the joint state pattern, and determine multiple migration patterns based on the multiple joint state migration paths. The health monitoring and analysis module is used to construct a state migration fingerprint for each migration pattern based on multiple behavioral change sequences and physiological change sequences. After collecting the target behavioral data and target physiological data of the user to be monitored, the module performs health management monitoring on the user to be monitored based on multiple state migration fingerprints, and obtains the health monitoring results of the user to be monitored during the virtual reality scene experience.

[0052] The above are merely specific embodiments of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art. Parts not described in detail in this specification are prior art known to those skilled in the art.

Claims

1. A health management and monitoring method based on virtual reality, characterized in that, include: Acquire behavioral and physiological data of multiple users during their virtual reality experience, construct multiple behavioral and physiological change sequences for each user, perform behavioral and physiological mutation detection for each user, and identify multiple behavioral and physiological mutation points for each user. Multiple behavioral and physiological mutation points of users are fused to determine multiple joint change points, and multiple joint state sequences of users are extracted from multiple behavioral and physiological change sequences based on multiple joint change points. Perform joint state pattern recognition on multiple joint state sequences to generate multiple joint state patterns for multiple users. Extract the joint state transition path for each user based on the joint state pattern and determine multiple transition patterns based on the multiple joint state transition paths. A state transition fingerprint is constructed for each transition mode based on multiple behavioral and physiological change sequences. After collecting the target behavioral and physiological data of the user to be monitored, health management monitoring is performed on the user to be monitored based on multiple state transition fingerprints to obtain the health monitoring results of the user to be monitored during the virtual reality scene experience.

2. The health management and monitoring method based on virtual reality according to claim 1, characterized in that, Identifying multiple behavioral and physiological mutation points for each user includes: Construct a behavior change matrix and a physiological change matrix for each user, and perform sliding window distribution difference analysis on the behavior change matrix and the physiological change matrix respectively. Calculate the sliding behavior difference of the behavior change matrix between any two adjacent sliding windows, and calculate the sliding physiological difference of the physiological change matrix between any two adjacent sliding windows. Based on the sliding behavior difference and the sliding physiological difference, determine multiple behavioral change points and physiological change points for the user.

3. The health management and monitoring method based on virtual reality according to claim 2, characterized in that, By fusing multiple behavioral and physiological mutation points of a user to identify multiple joint change points, and based on these joint change points, multiple joint state sequences of the user are extracted from multiple behavioral and physiological change sequences, including: The system fuses multiple behavioral and physiological mutation points for each user, including constructing a joint mutation curve for the user, performing a sliding scan on multiple mutation points in the joint mutation curve, and if the time difference between the current mutation point and the next mutation point is less than a preset tolerance threshold, then the current mutation point and the next mutation point are fused, including replacing the current mutation point and the next mutation point with the mean of the two mutation points, and outputting multiple joint change points after traversing multiple mutation points in the joint mutation curve. Multiple segment intervals of the user are constructed based on multiple joint change points. The user's behavior change matrix and physiological change matrix are extracted from multiple interval behavioral features in each segment interval, and the user's physiological change matrix and physiological features in each segment interval are extracted from multiple interval behavioral features and physiological features. A joint state sequence for each segment interval is constructed based on the multiple interval behavioral features and physiological features.

4. The health management and monitoring method based on virtual reality according to claim 3, characterized in that, Based on the joint state pattern, the joint state migration path for each user is extracted, and multiple migration patterns are determined based on the multiple joint state migration paths, including: Determine the joint state pattern of the user in each segment interval, and construct the joint state transition path of the user based on the joint state pattern of the user in multiple segment intervals; Extract multiple transition patterns from each joint state transition path, where each transition pattern includes two adjacent joint state patterns in the joint state transition path; For multiple joint state patterns, multiple joint state sequences are clustered to generate multiple joint state patterns.

5. A health management and monitoring method based on virtual reality according to claim 4, characterized in that, The state transition fingerprint for each migration pattern is constructed based on multiple behavioral and physiological change sequences, including: Determine multiple joint state migration paths to which each migration pattern belongs, and construct a migration sample library for each migration pattern. The migration sample library includes local behavior matrices and local physiological matrices extracted from multiple users containing the corresponding migration patterns. Construct the migration start state vector and migration end state vector for each user in the migration sample library with respect to the migration mode. Generate the migration progress curve for each user based on the migration start state vector and migration end state vector. This includes calculating the state migration distance based on the migration start state vector and migration end state vector, determining the user's joint migration state vector at any time, determining the user's displacement vector under the joint migration state vector based on the migration start state vector, constructing the user's migration direction vector with respect to the migration mode based on the migration start state vector and migration end state vector, calculating the projection of the user's displacement vector onto the migration direction vector, obtaining the user's migration progress value under the joint migration state vector, and constructing the user's migration progress curve based on the user's migration progress values ​​at multiple time points. The migration start feature, migration end feature, and migration path length feature of the user are extracted from the migration progress curve to generate the individual migration fingerprint of the user. The individual migration fingerprints of multiple users in the migration sample library are then merged to obtain the state migration fingerprint of the migration pattern.

6. A health management and monitoring method based on virtual reality according to claim 5, characterized in that, Health management monitoring of users based on multiple state transition fingerprints includes: Based on the target behavior data and target physiological data of the user to be monitored, behavioral mutation detection and physiological mutation detection are performed on the user to be monitored to determine multiple target joint change points of the user to be monitored, and multiple target joint state sequences of multiple users to be monitored are extracted based on the target joint change points. Multiple target joint state sequences are matched with multiple joint state patterns respectively. Based on the state pattern matching results, the target migration path of the user to be monitored is determined, and the target migration pattern to which the target migration path belongs is identified. Construct a target migration fingerprint for the user to be monitored along the target migration path. Based on the state migration fingerprint of the target migration pattern, perform deviation detection on the target migration fingerprint, determine the deviation parameters of the target migration fingerprint, and generate health monitoring results for the user to be monitored regarding target behavioral data and target physiological data.

7. A health management and monitoring system based on virtual reality, characterized in that, The system is used to implement the virtual reality-based health management and monitoring method according to any one of claims 1-6, comprising: The sample data acquisition and analysis module is used to acquire behavioral and physiological data of multiple users during the virtual reality scene experience, construct multiple behavioral and physiological change sequences for each user, perform behavioral and physiological mutation detection for each user, and determine multiple behavioral and physiological mutation points for each user. The joint state analysis module is used to fuse multiple behavioral and physiological mutation points of a user to determine multiple joint change points, and extract multiple joint state sequences of the user from multiple behavioral and physiological change sequences based on multiple joint change points; The migration pattern recognition module is used to perform joint state pattern recognition on multiple joint state sequences, generate multiple joint state patterns corresponding to multiple users, extract the joint state migration path of each user based on the joint state pattern, and determine multiple migration patterns based on the multiple joint state migration paths. The health monitoring and analysis module is used to construct a state migration fingerprint for each migration pattern based on multiple behavioral change sequences and physiological change sequences. After collecting the target behavioral data and target physiological data of the user to be monitored, the module performs health management monitoring on the user to be monitored based on multiple state migration fingerprints, and obtains the health monitoring results of the user to be monitored during the virtual reality scene experience.

Images