Virtual reality-based elderly group intelligent activity interaction method and system

Abstract

The invention discloses an elderly group intelligent activity interaction method and system based on virtual reality, and the method comprises the steps: obtaining health basic data and activity intention information of an elderly user, carrying out the embedded representation, and reading the health condition representation and activity intention representation of the elderly user; generating an activity item recommendation constraint according to the activity intention representation, and obtaining a recommended activity item and a recommended activity duration of the elderly user based on the activity recommendation constraint and the health condition representation; generating a user demand according to the motion feature flow, the recommended activity duration and the activity intention information, and converting the user demand into a virtual reality design element to construct a virtual scene; interaction feedback is obtained according to interaction element analysis of the virtual scene, the activity effect of the elderly user is evaluated, and self-adaptive activity item adjustment is carried out. The virtual reality technology is utilized to improve the activity participation power of the old people and increase the accessibility of activities in the mechanism to the old people, the activities of the old people in the mechanism are promoted, and sedentariness is reduced.

Description

Technical Field

[0001] The present invention relates to the field of virtual reality technology, and more specifically, to a method and system for elderly group intelligent activity interaction based on virtual reality. Background Art

[0002] Virtual reality technology can simulate various scenarios and activities, such as traveling, visiting museums, and taking courses, providing a variety of entertainment options for the elderly. It can also provide services such as gaming and entertainment, which helps to enhance the physical and mental health of the elderly. Therefore, how to use virtual reality to diversify the activities of the elderly and reduce the risk of long periods of sitting is an urgent problem to be solved. Summary of the invention

[0003] In order to solve the above technical problems, the present invention proposes a method and system for intelligent activity interaction for the elderly based on virtual reality. The purpose is to solve the problem of lack of diversity in entertainment activities in existing nursing homes by constructing virtual reality scenes to simulate scenes for physical activities and group activities of the elderly in nursing homes.

[0004] The first aspect of the present invention provides a method for elderly group intelligent activity interaction based on virtual reality, comprising:

[0005] Obtaining basic health data and activity intention information of the elderly user, preprocessing the basic health data and activity intention information, and reading the health status representation and activity intention representation of the elderly user;

[0006] generating activity item recommendation constraints according to the activity intention expression, obtaining recommended activity items for elderly users based on the activity recommendation constraints and the health status expression, and setting recommended activity durations for the recommended activity items;

[0007] Extracting motion features from the recommended activity items to construct a motion feature stream, generating user needs in combination with activity selection feedback, converting the user needs into virtual reality design elements, and constructing a virtual scene based on the design elements;

[0008] The interactive elements in the virtual scene are read, and interactive feedback is obtained according to the interactive element analysis. The activity effect of the elderly user is evaluated through the interactive feedback, and the activity items are adjusted adaptively.

[0009] In this solution, the basic health data and activity intention information are preprocessed to read the health status representation and activity intention representation of the elderly user, specifically:

[0010] Read the physical examination data of the elderly user and the monitoring data of the wearable health monitoring device as the basic health data of the elderly user, obtain the questionnaire survey data of the elderly user and the interactive activity items within the preset time period as the activity intention information of the elderly user;

[0011] Preprocessing the basic health data, extracting multidimensional features from the preprocessed basic health data using two parallel data processing branches, extracting low-dimensional features of the basic health data in the main data processing branch, importing the low-dimensional features into the convolutional neural network channel, and introducing channel attention to extract high-dimensional features, and obtaining an output vector of the main data processing branch;

[0012] In the side data processing branch, the pooling window is used to fill the average value of the low-dimensional features in the window into the corresponding position of the output feature vector according to the pooling step size to obtain the output vector of the side data processing branch;

[0013] The output vector of the main data processing branch and the output vector of the side data processing branch are merged to obtain multi-dimensional features and generate a health status representation of the elderly user;

[0014] The questionnaire data and interactive activity item lists of elderly users are preprocessed to extract the corresponding word vectors. The obtained word vectors are used for bidirectional feature extraction using the Bi-GRU model, and a multi-head attention mechanism is introduced to weight the hidden states to obtain the activity intention representation containing the activity interests of elderly users.

[0015] In this solution, activity item recommendation constraints are generated according to the activity intention expression, and recommended activity items for elderly users are obtained based on the activity recommendation constraints and the health status expression, specifically:

[0016] Based on big data and data analysis methods, a preset number of activity project categories that are of interest to the elderly group are obtained, activity project recommendation constraints are generated according to the activity intention expression, activity project categories are preliminarily selected using the activity project recommendation constraints, and primary weights are set for the preliminarily selected activity projects;

[0017] Retrieve and obtain motion data in rehabilitation training instances and daily care instances according to the health status representation of the elderly user, calculate Pearson correlation parameters with each activity item category according to the motion data, and filter relevant activity item categories that meet preset thresholds through the Pearson correlation parameters;

[0018] Aggregate the primary activity item categories and the related activity item categories to generate a set of activity items to be selected, perform a graph representation of the health status representation of the elderly user and the set of activity items to be selected, and perform a random walk based on the graph representation with the node of the health status of the elderly user as the starting point;

[0019] Obtain a meta-path containing several activity items obtained by random walk, use a graph attention network to learn and represent the meta-path, calculate the weight of each activity item node in the meta-path, and represent the importance of the node to the health status of the elderly user;

[0020] The meta-path with the highest total weight is selected, and the activity project nodes therein are extracted as the recommended activity projects for elderly users.

[0021] In this plan, the recommended activity duration of the recommended activity project is set as follows:

[0022] Obtaining rehabilitation training instances and daily care instances corresponding to the recommended activity items, analyzing the corresponding health status intervals and the corresponding activity duration intervals in the rehabilitation training instances, and constructing a mapping relationship between the health status intervals and the activity duration intervals;

[0023] The health status representation of the elderly user is obtained, the corresponding activity duration sub-interval is located based on the mapping relationship, and a nonlinear adaptive factor is introduced based on the activity duration sub-interval to obtain the recommended activity duration function of the elderly user to determine the recommended activity duration of the recommended activity item.

[0024] In this solution, the motion features in the recommended activity items are extracted to construct a motion feature stream, user needs are generated in combination with activity selection feedback, the user needs are converted into virtual reality design elements, and a virtual scene is constructed based on the design elements, specifically:

[0025] Decompose the motion features according to the rehabilitation training instances and daily care instances corresponding to the recommended activity items, add timestamps to the motion features according to the activity duration in the instances, connect the motion features in series according to the timestamps, and construct a global motion feature stream by aggregating the series order of different instances;

[0026] Obtaining activity selection feedback of elderly users for each recommended activity item and its global motion feature stream, reading the activity item category and the selected motion feature in the activity selection feedback, analyzing the richness of the motion features selected by the elderly users, adjusting the selected motion features based on the richness, and constructing the motion feature stream in combination with the recommended activity duration;

[0027] Generating user needs according to the recommended activity item categories selected by the elderly user and the motion feature stream, and extracting visual elements, auditory elements, human-computer elements and interactive elements of the activity virtual scene based on the user needs;

[0028] The virtual scene is initialized by the obtained design elements, and the feedback from elderly users is used to optimize the virtual scene.

[0029] In this solution, the interactive elements in the virtual scene are read, and the interactive feedback is obtained according to the interactive element analysis. The activity effect of the elderly user is evaluated through the interactive feedback, specifically:

[0030] Obtain interactive feedback of elderly users corresponding to interactive actions when they perform interactive elements in virtual scenes, and obtain index parameters of each motion feature in the motion feature flow from four aspects: motion speed, motion completion, motion smoothness and motion accuracy;

[0031] Constructing an indicator parameter matrix according to the indicator parameters, and constructing a threshold matrix according to a preset evaluation indicator threshold, and obtaining a residual vector between the indicator parameter matrix and the threshold matrix;

[0032] The weight is set by the feature step length occupied by each motion feature in the motion feature stream, and the residual vector of the motion is weighted summed to obtain the evaluation result of the activity effect of the elderly user.

[0033] In this plan, adaptive activity item adjustments are carried out, specifically:

[0034] According to the interaction feedback, the interaction elements that the elderly user does not meet the target are obtained and marked, and the position deviation corresponding to the marked interaction element is read, and the position deviation is divided according to the preset deviation threshold. If it is not greater than, the target position of the interaction element is reset according to the current reachable situation of the elderly user's action;

[0035] If it is greater than, the motion characteristics of the interaction elements corresponding to the position deviation are taken as the dependent variable, and the interaction elements that meet the action requirements are taken as the retained motion characteristics set. The motion characteristics are decomposed according to the rehabilitation training examples and daily care examples corresponding to the recommended activity items to construct the original motion characteristics set;

[0036] Extracting and retaining the action-reachable features of the elderly users in the motion feature set, constructing a retrieval tag according to the action-reachable features, and screening the action features that meet the similarity standard in the original motion feature set to generate the action feature set to be introduced;

[0037] The mRMR algorithm is expanded using the Pearson correlation coefficient to obtain the motion feature with the greatest correlation with the dependent variable in the set of motion features to be introduced, and the fitness ranking of the introduced motion features is obtained based on the principle that the motion feature has the greatest correlation with the dependent variable and the redundancy in the set of retained motion features is the smallest;

[0038] The motion features with the highest adaptability are introduced to replace the motion features of the interactive elements corresponding to the position deviation, so as to realize the adjustment of the activity items.

[0039] The second aspect of the present invention provides a virtual reality-based elderly group intelligent activity interaction system, including a data acquisition module, an activity recommendation module, a virtual scene module, an activity evaluation module and an evaluation result output module;

[0040] The data collection module is used to collect basic health data and activity intention information of elderly users and perform preprocessing to obtain health status representation and activity intention representation of elderly users;

[0041] The activity recommendation module obtains recommended activity items for the elderly user based on the activity interest constraints and health status of the elderly user, and obtains recommended activity duration of the recommended activity items;

[0042] The virtual scene module is used to determine user needs according to motion characteristics and activity intention information, extract virtual reality design elements according to user needs, and construct a virtual scene through the design elements;

[0043] The activity evaluation module reads the interactive feedback of the interactive elements in the virtual scene and evaluates the activity effect of the elderly user through the interactive feedback;

[0044] The evaluation result output module is used to output the evaluation data of the elderly user's activity process and perform data visualization.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] The present invention uses virtual reality technology to provide virtual scenes for physical activities of the elderly in nursing homes, including virtual scenes for group activities. Starting from the health status of the elderly users, activity projects with a high degree of adaptability to the health status and activity interests of the elderly users are generated, thereby increasing the motivation of the elderly to participate in activities and increasing the accessibility of activities within the institution to the elderly, thereby promoting activities of the elderly in the institution and reducing long periods of sitting. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the drawings required for use in the embodiments or exemplary descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to the drawings without paying creative work.

[0048] Figure 1 A flow chart of the elderly group intelligent activity interaction method based on virtual reality of the present invention is shown;

[0049] Figure 2 A flowchart of obtaining recommended activity items for elderly users according to an embodiment of the present invention is shown;

[0050] Figure 3 A flow chart showing the effect of an activity of an elderly user evaluated according to an embodiment of the present invention is shown;

[0051] Figure 4 The block diagram of the elderly group intelligent activity interaction system based on virtual reality of the present invention is shown. DETAILED DESCRIPTION

[0052] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0053] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0054] Figure 1 The flowchart of the virtual reality-based elderly group intelligent activity interaction method of the present invention is shown.

[0055] like Figure 1 As shown, in the first embodiment of the present invention, a method for elderly group intelligent activity interaction based on virtual reality is provided, comprising:

[0056] S102, obtaining basic health data and activity intention information of the elderly user, preprocessing the basic health data and activity intention information, and reading the health status representation and activity intention representation of the elderly user;

[0057] S104, generating activity item recommendation constraints according to the activity intention expression, obtaining recommended activity items for the elderly user based on the activity recommendation constraints and the health status expression, and setting recommended activity durations for the recommended activity items;

[0058] S106, extracting motion features from the recommended activity items to construct a motion feature stream, generating user needs in combination with activity selection feedback, converting the user needs into virtual reality design elements, and constructing a virtual scene based on the design elements;

[0059] S108, reading the interactive elements in the virtual scene, acquiring interactive feedback according to the interactive element analysis, evaluating the activity effect of the elderly user through the interactive feedback, and making adaptive activity item adjustments.

[0060] It should be noted that different preset medical data sources are accessed, the physical examination data of the elderly user within the preset time and the monitoring data of the wearable health monitoring device are read as the basic health data of the elderly user, and the questionnaire survey data of the elderly user and the interactive activity items within the preset time period are obtained as the activity intention information of the elderly user; the basic health data is preprocessed, and the preprocessed basic health data is subjected to multidimensional feature extraction using two parallel data processing branches, and the low-dimensional features of the basic health data are extracted in the main data processing branch and the low-dimensional features are introduced into the convolutional neural network channel, and the channel attention is introduced to extract high-dimensional features, and the output vector of the main data processing branch is obtained; in the side data processing branch, the average value of the low-dimensional features in the window is filled into the corresponding position of the output feature vector according to the pooling step size using the pooling window, and the output vector of the side data processing branch is obtained; the features of the basic health data extracted by the two branches include both deep-level high-dimensional complex features and low-dimensional feature information closely related to the original data, which is represented by the features of the rich basic health data. The output vector of the main data processing branch and the output vector of the side data processing branch are fused to obtain multidimensional features and generate the health status representation of the elderly user.

[0061] The questionnaire data and interactive activity list of elderly users are preprocessed to extract the corresponding word vectors. The obtained word vectors are used for bidirectional feature extraction using the Bi-GRU model. The time-related features of activity items of interest to elderly users are learned through the GRU network. A multi-head attention mechanism is introduced to weight the hidden state, and the fully connected layer is used to obtain the activity intention representation containing the activity interests of elderly users.

[0062] Figure 2 A flow chart of obtaining recommended activity items for elderly users according to an embodiment of the present invention is shown.

[0063] According to an embodiment of the present invention, activity item recommendation constraints are generated according to the activity intention expression, and recommended activity items for elderly users are obtained based on the activity recommendation constraints and the health status expression, specifically:

[0064] S202, based on big data and data analysis methods, obtaining a preset number of activity project categories that the elderly group is interested in, generating activity project recommendation constraints according to the activity intention expression, using the activity project recommendation constraints to preliminarily select activity project categories, and setting primary weights for the preliminarily selected activity projects;

[0065] S204, retrieving and acquiring motion data in rehabilitation training instances and daily care instances according to the health status representation of the elderly user, calculating Pearson correlation parameters with each activity item category according to the motion data, and screening relevant activity item categories that meet a preset threshold through the Pearson correlation parameters;

[0066] S206, aggregating the primary activity item categories and the related activity item categories to generate a set of activity items to be selected, performing a graph representation of the health status representation of the elderly user and the set of activity items to be selected, and performing a random walk based on the graph representation with the node of the health status of the elderly user as the starting point;

[0067] S208, obtaining a meta-path including several activity items obtained by random walking, using a graph attention network to learn and represent the meta-path, calculating the weight of each activity item node in the meta-path, and representing the importance of the node to the health status of the elderly user;

[0068] S210 , screening the meta-path with the highest total weight, and extracting the activity project nodes therein as recommended activity projects for elderly users.

[0069] It should be noted that the meta-path is obtained through Metapath random walk and other methods, and the meta-path is learned and represented using the graph attention network. The importance of each activity project node in the meta-path to the head node (the health status node of the elderly user) is learned according to the graph convolution and attention mechanism. The weight of each activity project node is obtained according to the importance. If each activity project node in the meta-path has a primary weight, the obtained weight is combined with the primary weight to obtain the total weight of all activity project nodes in the meta-path. After sorting the meta-path using the total weight, the activity project node in the meta-path with the highest total weight is obtained as the recommended activity project for the elderly user.

[0070] It should be noted that the rehabilitation training examples and daily care examples corresponding to the recommended activity items are obtained, the corresponding health status intervals and the activity duration intervals of a single activity in the rehabilitation training examples are analyzed, and a mapping relationship between the health status intervals and the activity duration intervals is constructed, and each health status maps several single activity durations. The health status representation of the elderly user is obtained, the corresponding activity duration sub-interval is located based on the mapping relationship, and a nonlinear adaptive factor is introduced based on the activity duration sub-interval to obtain the recommended activity duration function of the elderly user, so as to determine the recommended activity duration of the recommended activity items. The recommended activity duration function λ is expressed as: λ c , z The upper and lower limits of the activity duration subinterval of the distribution, a represents a fixed constant, gammainv represents an incomplete function used to update the activity duration function, and k represents the health status.

[0071] It should be noted that the motion features are decomposed according to the rehabilitation training instances and daily care instances corresponding to the recommended activity items, timestamps are added to the motion features according to the activity duration in the instances, the motion features are connected in series according to the timestamps, and a global motion feature stream is constructed by aggregating the series sequence of different instances; the activity selection feedback of the elderly users for each recommended activity item and its global motion feature stream is obtained, the activity item category and the selected motion features in the activity selection feedback are read, and the richness of the motion features selected by the elderly users is analyzed. The richness is obtained through the correlation of the motion features selected by the elderly users, and the selected motion features are adjusted based on the richness. When the action features selected by the elderly users are relatively single, the action feature categories are adaptively enriched. In addition, when the difficulty coefficient of the action features selected by the elderly users is analyzed, the action features of high difficulty levels are appropriately increased to increase the activity amount of the elderly users, and the motion feature stream is constructed in combination with the recommended activity duration.

[0072] Generate user needs based on the recommended activity item categories selected by the elderly user and the motion feature flow, and extract the visual elements, auditory elements, human-machine elements and interactive elements of the activity virtual scene based on the user needs; the visual elements in the virtual scene need to have obvious visibility and recognition. Avoid the elderly user from making inappropriate distinctions, which may lead to errors in activity evaluation. Improve the visibility and recognition of elements in the virtual scene by increasing the brightness and contrast of the virtual scene. For the auditory elements in the virtual scene, clear sound prompts are required to ensure the clarity of the sound prompts. In addition, the brisk background sound effects of the virtual scene can be set according to the music interests of the elderly user. For the human-machine elements in the virtual scene, consider the ergonomic issues related to size, position, and angle to ensure the safety and comfort of the elderly user and the accessibility of the elderly user's actions. For the interactive elements in the virtual scene, it is necessary to meet the understanding ability of the elderly user, add more spatial guidance and prompts to the interactive elements, reduce the complexity of the interaction, and ensure that the elderly user can understand clearly and quickly.

[0073] Figure 3 A flow chart of evaluating the activity effects of elderly users according to an embodiment of the present invention is shown.

[0074] According to an embodiment of the present invention, the interactive elements in the virtual scene are read, the interactive feedback is obtained according to the interactive element analysis, and the activity effect of the elderly user is evaluated through the interactive feedback, specifically:

[0075] S302, obtaining interactive feedback corresponding to interactive actions when the elderly user performs interactive elements in the virtual scene, and obtaining index parameters of each motion feature in the motion feature flow from four aspects: motion speed, motion completion, motion smoothness, and motion accuracy;

[0076] S304, constructing an indicator parameter matrix according to the indicator parameters, and constructing a threshold matrix according to a preset evaluation indicator threshold, and obtaining a residual vector between the indicator parameter matrix and the threshold matrix;

[0077] S306, setting weights according to the feature step lengths occupied by various motion features in the motion feature stream, performing weighted summation on the motion residual vectors, and obtaining evaluation results of the activity effects of the elderly user.

[0078] It should be noted that the motion speed index includes the average speed and maximum speed of elderly users in the movement of interactive elements; the motion completion index obtains the number of times the user completes the interactive elements in the activity and the difficulty coefficient of the interactive elements. When the difficulty coefficient is large, the threshold of the motion completion index is appropriately lowered; the motion smoothness is obtained by calculating the rate of change of acceleration of elderly users in the movement of interactive elements; the motion accuracy includes motion trajectory deviation and target position deviation, which are used to measure the activity quality of elderly users.

[0079] It should be noted that, based on the interaction feedback, the interaction elements that do not meet the standards for elderly users are obtained and marked, and the position deviations corresponding to the marked interaction elements are read. The position deviations are divided according to preset deviation thresholds of different levels, and the primary deviation threshold is set according to the adaptive interaction difficulty adjustment, and action features that are not greater than the primary deviation threshold are regarded as meeting the standards. An advanced threshold is set on the basis of the primary threshold, and the action feature is adjusted and judged according to the advanced threshold. If the primary threshold is not greater than the advanced threshold, the target position of the interactive element is reset according to the current reachable action of the elderly user; if the primary threshold is greater than the advanced threshold, the motion feature of the interactive element corresponding to the position deviation is used as the dependent variable, and the interactive element that meets the action requirements is used as the retained motion feature set, and the motion features are decomposed according to the rehabilitation training examples and daily care examples corresponding to the recommended activity items to construct an original motion feature set; the reachable action features of the elderly user in the retained motion feature set are extracted, and a retrieval tag is constructed according to the reachable action features, and the action features that meet the similarity standard in the original motion feature set are screened to generate a set of action features to be introduced; the Pearson correlation coefficient is used to expand the mRMR algorithm, and the motion feature with the greatest correlation with the dependent variable is obtained in the set of action features to be introduced, and the fitness ranking of the introduced motion features is obtained based on the principle that the correlation between the motion feature and the dependent variable is the greatest and the redundancy in the retained motion feature set is the smallest; the motion feature with the highest fitness is introduced to replace the motion feature of the interactive element corresponding to the position deviation, so as to achieve the adjustment of the activity items.

[0080] In addition, in the simulation scenario, substandard interactive elements are obtained based on position deviations, and spatial position distribution characteristics of the substandard interactive elements are generated. A substandard interaction heat map is generated based on the spatial position distribution characteristics. Based on the substandard interaction heat map, location areas with weak control capabilities during the activities of elderly users are obtained, and the weak control capability location areas are used to provide feedback to the virtual scenes of the elderly users, thereby increasing the number and difficulty of interactive element tasks in the weak control capability location areas, and achieving targeted activity training for elderly users.

[0081] Figure 4 The block diagram of the elderly group intelligent activity interaction system based on virtual reality of the present invention is shown.

[0082] The second embodiment of the present invention provides a virtual reality-based elderly group intelligent activity interaction system 4, including a data collection module 41, an activity recommendation module 42, a virtual scene module 43, an activity evaluation module 44 and an evaluation result output module 45;

[0083] The data collection module is used to collect basic health data and activity intention information of elderly users and perform preprocessing to obtain health status representation and activity intention representation of elderly users;

[0084] The activity recommendation module obtains recommended activity items for the elderly user based on the activity interest constraints and health status of the elderly user, and obtains recommended activity duration of the recommended activity items;

[0085] The virtual scene module is used to determine user needs according to motion characteristics and activity intention information, extract virtual reality design elements according to user needs, and construct a virtual scene through the design elements;

[0086] The activity evaluation module reads the interactive feedback of the interactive elements in the virtual scene and evaluates the activity effect of the elderly user through the interactive feedback;

[0087] The evaluation result output module is used to output the evaluation data of the elderly user's activity process and perform data visualization.

[0088] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0089] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0090] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0091] Those skilled in the art can understand that: all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), disks or optical disks, and other media that can store program codes.

[0092] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0093] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for elderly group intelligent activity interaction based on virtual reality, characterized in that: The following steps are involved: Obtaining basic health data and activity intention information of the elderly user, preprocessing the basic health data and activity intention information, and reading the health status representation and activity intention representation of the elderly user; generating activity item recommendation constraints according to the activity intention expression, obtaining recommended activity items for elderly users based on the activity recommendation constraints and the health status expression, and setting recommended activity durations for the recommended activity items; Extracting motion features from the recommended activity items to construct a motion feature stream, generating user needs in combination with activity selection feedback, converting the user needs into virtual reality design elements, and constructing a virtual scene based on the design elements; The interactive elements in the virtual scene are read, and interactive feedback is obtained according to the interactive element analysis. The activity effect of the elderly user is evaluated through the interactive feedback, and the activity items are adjusted adaptively.

2. The method for elderly group intelligent activity interaction based on virtual reality according to claim 1 is characterized in that: The basic health data and activity intention information are preprocessed to read the health status representation and activity intention representation of the elderly user, specifically: Read the physical examination data of the elderly user and the monitoring data of the wearable health monitoring device as the basic health data of the elderly user, obtain the questionnaire survey data of the elderly user and the interactive activity items within the preset time period as the activity intention information of the elderly user; Preprocessing the basic health data, extracting multidimensional features from the preprocessed basic health data using two parallel data processing branches, extracting low-dimensional features of the basic health data in the main data processing branch and importing the low-dimensional features into the convolutional neural network channel, and introducing channel attention to extract high-dimensional features, and obtaining an output vector of the main data processing branch; In the side data processing branch, the pooling window is used to fill the average value of the low-dimensional features in the window into the corresponding position of the output feature vector according to the pooling step size to obtain the output vector of the side data processing branch; The output vector of the main data processing branch and the output vector of the side data processing branch are merged to obtain multi-dimensional features and generate a health status representation of the elderly user; The questionnaire data and interactive activity item lists of elderly users are preprocessed to extract the corresponding word vectors. The obtained word vectors are used for bidirectional feature extraction using the Bi-GRU model, and a multi-head attention mechanism is introduced to weight the hidden states to obtain the activity intention representation containing the activity interests of elderly users.

3. The method for elderly group intelligent activity interaction based on virtual reality according to claim 1 is characterized in that: Generate activity item recommendation constraints according to the activity intention expression, and obtain recommended activity items for elderly users based on the activity recommendation constraints and the health status expression, specifically: Based on big data and data analysis methods, a preset number of activity project categories that are of interest to the elderly group are obtained, activity project recommendation constraints are generated according to the activity intention expression, activity project categories are preliminarily selected using the activity project recommendation constraints, and primary weights are set for the preliminarily selected activity projects; Retrieve and obtain action data in rehabilitation training instances and daily care instances according to the health status representation of the elderly user, calculate the Pearson correlation parameters with each activity item category according to the action data, and filter the relevant activity item categories that meet the preset threshold value through the Pearson correlation parameters; Aggregate the primary activity item categories and the related activity item categories to generate a set of activity items to be selected, perform a graph representation of the health status representation of the elderly user and the set of activity items to be selected, and perform a random walk based on the graph representation with the node of the health status of the elderly user as the starting point; Obtain a meta-path containing several activity items obtained by random walk, use a graph attention network to learn and represent the meta-path, calculate the weight of each activity item node in the meta-path, and represent the importance of the node to the health status of the elderly user; The meta-path with the highest total weight is selected, and the activity project nodes therein are extracted as the recommended activity projects for elderly users.

4. The method for elderly group intelligent activity interaction based on virtual reality according to claim 1 is characterized in that: Set the recommended activity duration for the recommended activity project, specifically: Obtaining rehabilitation training instances and daily care instances corresponding to the recommended activity items, analyzing the corresponding health status intervals and the corresponding activity duration intervals in the rehabilitation training instances, and constructing a mapping relationship between the health status intervals and the activity duration intervals; The health status representation of the elderly user is obtained, the corresponding activity duration sub-interval is located based on the mapping relationship, and a nonlinear adaptive factor is introduced based on the activity duration sub-interval to obtain the recommended activity duration function of the elderly user to determine the recommended activity duration of the recommended activity item.

5. The method for elderly group intelligent activity interaction based on virtual reality according to claim 1 is characterized in that: Extract the motion features in the recommended activity items to construct a motion feature stream, generate user needs in combination with activity selection feedback, convert the user needs into virtual reality design elements, and construct a virtual scene based on the design elements, specifically: Decompose the motion features according to the rehabilitation training instances and daily care instances corresponding to the recommended activity items, add timestamps to the motion features according to the activity duration in the instances, connect the motion features in series according to the timestamps, and construct a global motion feature stream by aggregating the series order of different instances; Obtaining activity selection feedback of elderly users for each recommended activity item and its global motion feature stream, reading the activity item category and the selected motion feature in the activity selection feedback, analyzing the richness of the motion features selected by the elderly users, adjusting the selected motion features based on the richness, and constructing the motion feature stream in combination with the recommended activity duration; Generating user needs according to the recommended activity item categories selected by the elderly user and the motion feature stream, and extracting visual elements, auditory elements, human-computer elements and interactive elements of the activity virtual scene based on the user needs; The virtual scene is initialized by the obtained design elements, and the feedback from elderly users is used to optimize the virtual scene.

6. The method for elderly group intelligent activity interaction based on virtual reality according to claim 1 is characterized in that: Read the interactive elements in the virtual scene, obtain interactive feedback according to the interactive element analysis, and evaluate the activity effect of the elderly user through the interactive feedback, specifically: Obtain interactive feedback of elderly users corresponding to interactive actions when they perform interactive elements in virtual scenes, and obtain index parameters of each motion feature in the motion feature flow from four aspects: motion speed, motion completion, motion smoothness and motion accuracy; Constructing an indicator parameter matrix according to the indicator parameters, and constructing a threshold matrix according to a preset evaluation indicator threshold, and obtaining a residual vector between the indicator parameter matrix and the threshold matrix; The weight is set by the feature step length occupied by each motion feature in the motion feature stream, and the residual vector of the motion is weighted summed to obtain the evaluation result of the activity effect of the elderly user.

7. The method for elderly group intelligent activity interaction based on virtual reality according to claim 1 is characterized in that: Carry out adaptive activity project adjustments, specifically: According to the interaction feedback, the interaction elements that the elderly user does not meet the target are obtained and marked, and the position deviation corresponding to the marked interaction element is read, and the position deviation is divided according to the preset deviation threshold. If it is not greater than, the target position of the interaction element is reset according to the current reachable situation of the elderly user's action; If it is greater than, the motion characteristics of the interaction elements corresponding to the position deviation are taken as the dependent variable, and the interaction elements that meet the action requirements are taken as the retained motion characteristics set. The motion characteristics are decomposed according to the rehabilitation training examples and daily care examples corresponding to the recommended activity items to construct the original motion characteristics set; Extracting and retaining the action-reachable features of the elderly users in the motion feature set, constructing a retrieval tag according to the action-reachable features, and screening the action features that meet the similarity standard in the original motion feature set to generate the action feature set to be introduced; The mRMR algorithm is expanded using the Pearson correlation coefficient to obtain the motion feature with the greatest correlation with the dependent variable in the set of motion features to be introduced, and the fitness ranking of the introduced motion features is obtained based on the principle that the motion feature has the greatest correlation with the dependent variable and the redundancy in the set of retained motion features is the smallest; The motion features with the highest adaptability are introduced to replace the motion features of the interactive elements corresponding to the position deviation, so as to realize the adjustment of the activity items.

8. A virtual reality-based elderly group intelligent activity interaction system, characterized in that: It includes data collection module, activity recommendation module, virtual scene module, activity evaluation module and evaluation result output module; The data collection module is used to collect basic health data and activity intention information of elderly users and perform preprocessing to obtain health status representation and activity intention representation of elderly users; The activity recommendation module obtains recommended activity items for the elderly user based on the activity interest constraints and health status of the elderly user, and obtains recommended activity duration of the recommended activity items; The virtual scene module is used to determine user needs according to motion characteristics and activity selection feedback, extract virtual reality design elements according to user needs, and construct a virtual scene through the design elements; The activity evaluation module reads the interactive feedback of the interactive elements in the virtual scene and evaluates the activity effect of the elderly user through the interactive feedback; The evaluation result output module is used to output the evaluation data of the elderly user's activity process and perform data visualization.

Images