Individualized intervention strategy generation method driven by parent-child interaction type motion data

Abstract

The invention provides a personalized intervention strategy generation method driven by parent-child interaction type motion data. Through a multi-source fusion perception technology, motion tracking, bioelectricity sensing, environment sensing and social interaction data acquisition are integrated, and parent-child motion behaviors, physiological indexes, psychological emotion and environment data are comprehensively collected. A deep learning correlation analysis model is used for mining a complex nonlinear relationship among multi-source data, and parent-child groups are dynamically divided according to real-time data by means of a dynamic clustering evolutionary algorithm. And based on an evaluation result, customizing an exclusive exercise scheme by using a VR / AR technology, and constructing a real-time adjustment strategy of the emotion intelligent feedback system. Besides, through an adaptive learning adjustment engine and a community interaction and strategy sharing platform, an intervention strategy is dynamically optimized according to exercise data changes and user feedback, parent-child exercise experience is improved, and parent-child relationship and health level are promoted.

Description

Technical Field

[0001] The present invention relates to a method for generating personalized intervention strategies for parent-child interactive sports in the field of sports health management by using multi-source data fusion and deep learning algorithm technologies. Background Art

[0002] In today's society, with the continuous improvement of people's health awareness and the increasing emphasis on family relationships, parent-child interactive sports, as an activity form that can not only promote the physical health of family members but also effectively enhance parent-child emotional communication, have received extensive attention and favor. More and more parents realize that by participating in sports together, they can not only help children develop good exercise habits and promote their physical development but also deeply understand their children's inner world during the interaction process and establish a closer and more harmonious parent-child relationship.

[0003] Currently, there are many limitations in the guidance and intervention means for parent-child interactive sports on the market. Most of the existing exercise programs adopt general templates and fail to fully consider the significant individual differences among different parent-child combinations in terms of exercise ability, physical condition, interest preferences, and psychological and emotional needs. For parent-child families with stronger exercise ability, conventional low-intensity exercise programs cannot meet their needs for further improving their exercise level; while for parent-child with weaker exercise ability and special physical conditions, general high-intensity exercise plans may increase the risk of physical discomfort or even injury.

[0004] At the psychological and emotional level, each child and parent have different feelings, expectations, and preferences for the interaction methods of sports. Traditional intervention strategies lack accurate grasp and targeted guidance on these emotional factors and are difficult to effectively stimulate the enthusiasm and persistence of parents and children in participating in sports. This not only greatly reduces the exercise effect and fails to fully realize the potential value of parent-child sports in promoting health and enhancing emotions but also may lead to a sense of frustration during the parent-child sports process, thereby reducing the interest and participation in parent-child sports and being unfavorable to the cultivation of a long-term healthy lifestyle and the continuous improvement of parent-child relationships.

[0005] This patent aims to break through these limitations by using advanced multi-source data fusion technology to comprehensively and accurately collect various data during parent-child sports. By leveraging cutting-edge deep learning algorithms to deeply mine and analyze the data, it can keenly identify the uniqueness of different parent-child combinations in terms of sports ability, physical condition, interest preferences, and psychological and emotional needs. Based on this, a customized sports intervention strategy is developed for each parent-child group, achieving a leap from generalization to personalization. In terms of improving sports ability, for parent-child families at different levels, a suitable-intensity sports plan is accurately matched to help them steadily improve their sports level while reducing the risk of sports injuries. In terms of psychological and emotional guidance, with the help of an emotional intelligence feedback system, the emotional changes of parents and children are captured in real time, and the strategy is adjusted in a timely manner to provide appropriate emotional support and interactive guidance to both parents and children, fully stimulating their enthusiasm and persistence in participating in sports. The system constructed by this patent is committed to providing all-round, personalized, and dynamically optimized intervention strategies for parent-child sports, maximizing the potential of parent-child sports in promoting health and enhancing emotions, and effectively making up for the deficiencies of existing technologies. Summary of the Invention

[0006] The summary of the invention elaborates in detail the specific processes, technical key points, and innovations in data collection, analysis, intervention strategy generation, and dynamic adjustment in the patent, highlighting its advantages compared to traditional methods.

[0007] The innovative data collection methods include: Multi-source fusion perception technology: The present invention innovatively uses multi-source fusion perception technology to collect motion data. Motion tracking devices such as smart bracelets and sports watches are used to accurately collect basic motion behavior data such as the number of steps, distance, duration, speed, and exercise intensity during parent-child exercise at a frequency of 10 times per second. The exercise intensity is calculated by means of a specific kinematic algorithm using the data from the acceleration sensor and gyroscope built into the device. A wearable bioelectric sensor attached to the skin surface of the parent and child monitors the muscle electrical activity of the parent and child in real time at a frequency of 1000 Hz. The signals collected by the sensor are first pre-processed through filtering and amplification, and then an advanced muscle electrical activity analysis algorithm is used to evaluate the degree of muscle fatigue and the force application pattern. In an indoor exercise scenario, environmental sensors are deployed within 3 meters of the exercise area to collect environmental data such as temperature, humidity, and air quality during exercise at a frequency of 1 time per minute. Among them, the temperature measurement accuracy can reach ±0.5 °C, the humidity measurement accuracy can reach ±3% RH, and the air quality is measured by detecting the concentrations of PM2.5 and formaldehyde pollutants. In an outdoor scenario, an environmental sensor with waterproof, dustproof, and anti-interference capabilities is installed within 5 meters of the parent-child exercise path to ensure stable collection of environmental data. Social interaction data collection: A specially designed mobile application, after the user agrees to the privacy policy, enables the camera and microphone permissions. The application uses speech recognition technology to convert the voice communication content of the parent and child during exercise into text in real time with an accuracy of over 95%, and uses natural language processing technology for semantic analysis. A facial expression recognition algorithm based on deep learning can recognize 8 common expressions such as happy, frustrated, and excited, comprehensively reflecting the quality of emotional communication and interaction patterns between parents and children.

[0008] The described unique data analysis algorithms include: Deep learning association analysis model: A deep learning association analysis model is constructed, which adopts a multi-layer neural network structure. The input layer contains 100 nodes and receives multi-source data such as motion behavior data, physiological index data, psychological and emotional data, environmental data, and social interaction data. The model has 3 hidden layers with the number of nodes being 50, 30, and 20 respectively. The ReLU activation function is used for the operations in the hidden layers. After complex operations through multiple hidden layers, the 10 nodes in the output layer output the results of the association relationships between the data, automatically mining the complex non-linear relationships between the multi-source data. Dynamic clustering evolution algorithm: The dynamic clustering evolution algorithm is adopted, which is based on the K-Means clustering algorithm and combines the idea of genetic algorithm. Initially, 5 clustering centers are set. In each iteration process, the distances from each parent-child data point to each clustering center are calculated, and the data points are divided into the corresponding clusters according to the principle of the closest distance. The genetic algorithm is used to optimize the clustering centers, and through selection, crossover, and mutation operations, the dynamic update and optimization of the clustering centers are realized. According to the characteristics of the real-time updated motion data of the parent-child, including the motion ability (obtained by comprehensively evaluating the motion endurance, strength, and coordination indexes), motion preference (such as the preference degree for different motion types such as running, ball games, and parent-child games), and emotional interaction mode (analyzed according to the parent-child communication frequency and emotional expression intensity in the social interaction data), the clustering results are dynamically adjusted, and the parent-child are accurately divided into different group categories.

[0009] The innovative implementation of the personalized intervention strategy includes: Virtual Reality (VR) / Augmented Reality (AR) assisted exercise programs: Based on the results of the exercise ability assessment and combined with the exercise goals set by parents and children, a professional Virtual Reality (VR) / Augmented Reality (AR) technology development team is utilized to customize exclusive exercise programs for different groups of parents and children. During development, the differences in exercise abilities between parents and children are fully considered (such as dividing exercise abilities into three levels: low, medium, and high, and designing exercise sessions with different difficulties and intensities for different levels), goal setting (such as improving physical fitness, enhancing parent-child relationships, and cultivating an interest in exercise), and the adaptability of exercise scenarios (indoor scenarios focus on space utilization and safety guarantee, while outdoor scenarios emphasize integration with the natural environment). For the endurance-type parent-child group, a virtual marathon track running challenge is designed, and virtual track indicators (the track indicators are presented in the form of clear and eye-catching graphics and text, with a distance prompt accuracy of up to ±1 meter) and competitors (the virtual images of the competitors are customized according to the challenge difficulty and goals set by the parents and children, with different running speeds and endurance performances) are superimposed on the real scenario with the help of AR technology; for the strength-type parent-child group, a VR-based strength training game is developed, simulating a weightlifting competition scenario, with 5 weight levels, 3 different-shaped barbells, and realistic competition environment sound effects and audience interaction elements set in the virtual scenario. Emotional intelligence feedback system: An emotional intelligence feedback system is constructed, which has functional modules for emotional dictionary construction, semantic analysis, and emotional classification. The emotional dictionary is continuously expanded and optimized by collecting more than 5000 emotional words related to parent-child exercise and combining the emotional expression data in the actual exercise scenario. The semantic analysis module uses a natural language processing model based on the Transformer architecture to analyze the grammar and semantics of the speech communication text during the parent-child exercise process and extract key information. The emotional classification module uses deep learning algorithms such as Convolutional Neural Network (CNN) to classify the facial expressions and speech emotions of parents and children, and can accurately identify 8 emotional states. The system uses natural language processing and emotion recognition technologies to real-time monitor the emotional information conveyed by the speech and expressions of parents and children during exercise. When the system detects a change in emotional needs, it automatically adjusts the intervention strategy. For example, when it detects that the child is frustrated and struggling with the exercise difficulty, it pushes suggestions for reducing the difficulty and encouraging word templates to the parents through a pre-set rule base, and the encouraging word templates are customized according to factors such as the child's age and personality characteristics.

[0010] The innovative mechanism of the dynamic adjustment strategy includes: Adaptive learning adjustment engine: Deploy an adaptive learning adjustment engine, which is built-in with time series analysis algorithms (using the ARIMA(2,1,2) model) and machine learning classification algorithms (such as Support Vector Machine SVM). By monitoring the parent-child movement data in real time, analyze the data change situation at a frequency of 5 times per minute. When it is found that the movement ability has improved (such as an increase in endurance is manifested as an increase in the continuous 7-day movement duration by more than 10%, and the speed has increased by more than 5%), or the emotional needs have changed (such as in 5 consecutive movements, the child's positive emotional expression for a certain type of movement has decreased by more than 20%), the strategy adjustment mechanism is automatically triggered. If the improvement of parent-child movement endurance is monitored for two consecutive weeks, the engine automatically increases the movement duration and intensity according to the pre-set movement intensity adjustment model. The increase in movement duration is 15% of the current duration, and the intensity increase is reasonably adjusted according to the movement type and the physical conditions of the parent and child in accordance with professional sports training standards; if the interest of the parent and child in outdoor parent-child sports decreases, the engine recommends indoor parent-child sports projects and adjusts the sports plan based on historical data and preferences by analyzing the movement records and search preference data of users in the application. The recommended indoor parent-child sports projects include at least 3 projects that meet the interests and movement abilities of the parent and child.

[0011] Community interaction and strategy sharing platform: Build a community interaction and strategy sharing platform, which has functions such as user identity authentication (using multiple authentication methods based on mobile phone number, verification code, and face recognition), data security encryption (using the AES-256 encryption algorithm to encrypt and store and transmit user data), and personalized recommendation algorithm interface. Using big data analysis technology, mine and analyze the sports experience shared by parent-child families on the platform (presented in various forms such as text, pictures, and videos) and the effect feedback after implementing the intervention strategy (collected through questionnaires and user evaluations). According to the analysis results, recommend the successful intervention strategies of other similar parent-child families to families in need. The recommendation algorithm comprehensively considers factors in 8 dimensions including the movement ability, goals, emotional needs, and geographical location of the parent and child to achieve the continuous evolution and shared dissemination of personalized intervention strategies. At the same time, the platform optimizes the recommendation algorithm every 15 days. By collecting the usage feedback and effect evaluation of users on the recommended strategies, continuously adjust the algorithm parameters to improve the recommendation accuracy. Please check whether the details of the invention content meet your expectations. If you hope to further refine a certain technical point or adjust the expression of the process, you can let us know at any time.

[0012] The system integration and compatibility include: Multi-platform data interaction: The system constructed by the present invention has powerful multi-platform data interaction capabilities, and can seamlessly connect to mainstream sports health management platforms, smart device operating systems, and social software. It realizes data synchronization with common sports APPs, and the personalized exercise plans and exercise achievements generated by users in this system can be conveniently shared on social platforms, motivating parent-child families to continuously participate in sports, and at the same time attracting more families to pay attention to and adopt this personalized intervention strategy. In addition, the system can be connected to smart TV devices to provide richer exercise displays and interactive experiences for parent-child in the home scenario, such as presenting VR / AR exercise scenarios on a large screen to enhance the immersion. Device compatibility optimization: Considering the diversity of sports devices on the market, the present invention has deeply optimized device compatibility in the design. Whether it is common sports tracking device brands on the market or emerging wearable health monitoring devices, as long as they conform to the general data transmission protocol, they can be easily connected to this system. This feature ensures that parent-child families can make full use of existing device resources without having to purchase specific devices additionally, reducing the usage cost and improving the popularity and applicability of the present invention.

[0013] The privacy protection and data security enhancement include: Multi-level encryption system: In terms of data privacy protection, the present invention adopts a multi-level encryption system. Starting from the data collection end, all sensors and applications use the SSL / TLS encryption protocol when transmitting data to ensure the security of data during transmission. In the data storage link, in addition to encrypting and storing the entire database using the AES-256 encryption algorithm, key user identity information and sensitive physiological data are also separately encrypted. Homomorphic encryption is additionally performed on the physiological index data related to the medical health of parent-child, so that calculations and analyses can still be performed on the encrypted data, further ensuring data security. Strict access right management: A strict access right management mechanism is established, and only authorized system modules and specific administrator accounts can access and process user data. Different access levels are set for different types of data. Sports behavior data can be accessed by both authenticated parent and child and the data analysis module of the system; while sensitive data related to the mental health assessment of parent-child is only authorized to be accessed by professional psychologist accounts (after the explicit consent of parent and child) and senior administrator accounts with security audit functions. At the same time, the system records and audits all data access behaviors in detail so that anomalies can be traced and processed in a timely manner when they occur.

[0014] Adopting the above technical solutions can bring the following technical effects:

[0015] 1. Precision improvement of exercise effectiveness: Through multi-source fusion perception technology, exercise behavior data is collected at a high frequency of 10 times per second. Combined with a bioelectric sensor that monitors muscle electrical activity at 1000Hz, the parent-child exercise ability is accurately evaluated. Based on this, exclusive programs are customized for parent-child pairs with different exercise ability levels. For example, endurance-oriented parent-child pairs participate in virtual marathon track challenges with a distance prompt accuracy of ±1 meter to help them scientifically improve their endurance; strength-oriented parent-child pairs use VR for strength training games with 5 weight levels and 3 barbell shapes to specifically enhance strength. At the same time, the adaptive learning and adjustment engine analyzes data at a frequency of 5 times per minute. When the exercise endurance of parent-child pairs increases by more than 10% for 7 consecutive days, the exercise duration is automatically increased by 15% and a reasonable intensity is set to ensure continuous optimization of exercise effects.

[0016] 2. Deep promotion of emotional interaction: Social interaction data is collected using speech recognition (accuracy over 95%) and facial expression recognition (recognizing 8 common expressions) to comprehensively understand parent-child emotional communication. The emotional intelligence feedback system monitors emotional changes in real time. Once it detects that the child is frustrated and struggling with the exercise difficulty, it immediately pushes personalized encouragement word templates from a rule base constructed with more than 5000 emotional words, customized according to the child's age and personality, effectively relieving negative emotions and enhancing the emotional connection between parents and children. The community interaction and strategy sharing platform aggregates parent-child exercise experiences. The recommendation algorithm synthesizes 8-dimensional factors to provide reference for emotional interaction for parent-child families, promoting in-depth communication between parents and children during exercise and creating a harmonious atmosphere.

[0017] 3. Efficient implementation of long-term health management: The system comprehensively collects multi-source data on exercise, physiology, and environment to build a comprehensive health record. The multi-level encryption system and strict access permission management ensure data security and privacy. In cooperation with educational institutions, personalized exercise programs are integrated into physical education courses to help students develop exercise habits; in cooperation with medical institutions, exercise rehabilitation plans are developed for parent-child pairs with chronic diseases or in the recovery period, and real-time monitoring and adjustment are carried out. For children with obesity, a comprehensive exercise plan is designed and the compliance of rehabilitation is improved through the emotional intelligence feedback system. In addition, the platform optimizes the recommendation algorithm every 15 days to continuously provide scientific and effective long-term health management services for parent-child families. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flowchart of the dynamic clustering evolution algorithm of the present invention

[0019] Figure 2 It is an architecture diagram of the multi-source data acquisition system of the present invention DETAILED DESCRIPTION OF THE INVENTION

[0020] The following will be combined with the appended drawings in the embodiments of the present invention Figure 1 and Figure 2, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0021] Example 1: Technical points of daily exercise for ordinary healthy parent-child families: Using multi-source fusion perception technology, the motion tracking device collects motion step count, distance, and speed data 10 times per second, the bioelectric sensor monitors muscle electrical activity at 1000 Hz, and the environmental sensor collects environmental temperature and humidity data once per minute. The data association is mined through a deep learning correlation analysis model, and the dynamic clustering evolution algorithm classifies this family into an ordinary active parent-child group. Based on this, a customized daily exercise plan is developed for them, such as parent-child jogging, with virtual runway instructions superimposed on the real scene by means of AR technology, and the distance indication accuracy reaches ±1 meter. The adaptive learning and adjustment engine analyzes the data 5 times per minute. If it is found that the exercise endurance has increased by 10% for 7 consecutive days, the exercise duration is automatically increased by 15% and a reasonable intensity is added. Advantage: It accurately fits the exercise ability and needs of the family, avoiding excessive or insufficient exercise. The AR technology enhances the fun of exercise, and the real-time data monitoring and plan adjustment ensure the continuous optimization of the exercise effect, promoting the health of parents and children.

[0022] Example 2: Parent-child training for children with sports specialties: The motion tracking device and the bioelectric sensor continuously collect data, combined with the collection of social interaction data to understand the emotions and needs of the children during exercise. Since the children have sports specialties, they are clustered into a specialty parent-child group. A customized exclusive training plan is developed for them. For example, if the children are good at basketball, a VR-based parent-child basketball training game is designed, with 5 different difficulty levels and a variety of basketball training scenarios. The emotional intelligent feedback system monitors in real time. If the children are dissatisfied with the training difficulty, targeted encouragement and adjustment suggestions are pushed from a 5000-word emotional vocabulary rule base. Advantage: It gives full play to the sports specialties of the children, and the VR training game provides an immersive and personalized training environment. The emotional intelligent feedback system guarantees the enthusiasm of the children for training, improves the training effect, and enhances the interaction between parents and children in the cultivation of specialties.

[0023] Example 3: Key technical points of the enlightenment exercise for young children and parent-child families: The motion tracking device collects basic motion data, and the environmental sensor pays attention to the safety of the indoor environment. Understand the emotions of young children through voice recognition (accuracy exceeds 95%) and facial expression recognition (recognizes 8 common expressions). According to the characteristics of young children, they are clustered into early childhood enlightenment parent-child groups, and customized parent-child yoga enlightenment courses are provided, with low movement difficulty and strong fun. The community interaction and strategy sharing platform recommends sports experience and interaction methods suitable for young children, and the emotional intelligent feedback system soothes the emotions of young children in time. Advantages: Provide safe and appropriate sports enlightenment for young children, the social interaction data collection and emotional intelligent feedback system pay attention to the psychology of young children, and the community platform provides rich experience to help young children develop exercise habits and enhance parent-child relationships.

[0024] Example 4: Key points of parent-child weight loss exercise technology: Multi-source fusion sensing technology comprehensively collects exercise and physiological data, such as monitoring changes in muscle electrical activity related to fat metabolism through bioelectric sensors. Based on the data, the family is clustered into a parent-child group with weight loss needs. Customized weight loss exercise plans, including a combination of aerobic exercise and strength training, such as parent-child aerobics and simple parent-child strength training movements. The adaptive learning adjustment engine adjusts the intensity and duration of exercise based on weekly changes in weight and body fat percentage data, combined with exercise data. Advantages: Accurately target weight loss needs, and formulate scientific weight loss plans through comprehensive analysis of multi-source data. Real-time monitoring and adjustment ensure weight loss results, avoid health effects caused by improper exercise, and improve the success rate and experience of parent-child weight loss.

[0025] Example 5: Key technical points of low-intensity exercise for elderly parent-child families: Motion tracking devices collect low-intensity exercise data, and environmental sensors focus on monitoring environmental comfort and safety. Taking into account the physical condition of the elderly, they are clustered into elderly-care parent-child groups. Customized low-intensity exercise plans, such as parent-child slow walking and Tai Chi, use AR technology to display movement specifications and explanations. The emotional intelligent feedback system pays attention to the emotions of the elderly to avoid negative emotions due to exercise discomfort. Advantages: Fully consider the physical characteristics of the elderly and provide safe and appropriate low-intensity exercise plans. AR technology assists in ensuring the standardization of exercise movements, and the emotional intelligent feedback system provides emotional care for the elderly, promoting the health and emotional communication of elderly parent-child families.

[0026] Example 6: Parent-child families with special needs (such as children with mild autism) Technical points: Social interaction data collection focuses on the special behavioral manifestations of children. Motion tracking and bioelectric sensors collect motion data. Through analysis, they are clustered into parent-child groups with special needs. Customize specialized exercise programs, such as parent-child rhythm activities combining music and simple exercises, and use the emotional intelligence feedback system to adjust the activity rhythm and method in real time according to the child's emotional changes. The community interaction and strategy sharing platform shares the successful experiences and coping methods of similar families. Advantages: Provide personalized and targeted exercise interventions for parent-child families with special needs. Supported by the emotional intelligence feedback system and the community platform, it helps parents better guide their children to participate in exercises, promotes the physical and mental health development of children, and enhances the special emotional connection between parents and children.

[0027] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions in the specification are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of the present invention claimed is defined by the appended claims and their equivalents.

Claims

1. A method for generating personalized intervention strategies driven by parent-child interaction exercise data, characterized in that: The steps include: S1: Data collection: S101: Multi-source motion and environmental data collection: Using multi-source fusion sensing technology, with the help of motion tracking equipment, the basic motion behavior data of the number of steps, distance, duration, speed, and intensity of exercise in parent-child exercise is accurately collected at a frequency of 10 times per second. Among them, the intensity of exercise is calculated based on the data of the built-in accelerometer and gyroscope of the device through a specific kinematic algorithm. Through the wearable bioelectric sensor attached to the surface of the parent-child skin, the parent-child muscle electrical activity is monitored in real time at a frequency of 1000Hz. After signal filtering and amplification preprocessing, the advanced muscle electrical activity analysis algorithm is used to evaluate the degree of muscle fatigue and force mode. Using environmental sensors, in indoor sports scenes, they are deployed within 3 meters of the exercise area, and the temperature, humidity, and air quality environmental data of the environment during exercise are collected at a frequency of 1 time per minute. The temperature measurement accuracy can reach ±0.5℃, and the humidity measurement accuracy can reach ±3%RH. The air quality is measured by detecting the concentration of PM2.5 and formaldehyde pollutants. In outdoor scenarios, environmental sensors with waterproof, dustproof and anti-interference capabilities are used and installed within 5 meters of the parent-child exercise path to ensure stable collection of environmental data. S102: Social interaction data collection - With the help of a specially designed mobile application, after the user agrees to the privacy policy, the camera and microphone permissions are enabled. The application uses voice recognition technology to convert the voice communication content of parents and children during exercise into text in real time with an accuracy rate of more than 95%, and uses natural language processing technology for semantic analysis. Using a facial expression recognition algorithm based on deep learning, it can recognize 8 common expressions such as happiness, frustration, and excitement, which fully reflects the quality of emotional communication and interaction patterns between parents and children. S2: Data Analysis: S201: Data preprocessing and integration - All kinds of collected data are transmitted to high-performance data processing servers. Using data cleaning algorithms, outliers, duplicate values ​​and erroneous data are removed by setting a reasonable data range (such as the number of steps per day is between 1000-100000 steps as a reasonable range) and data fluctuation thresholds. Subsequently, the cleaned data is integrated into a unified database according to the parent-child combination. The database uses the relational database MySQL, which has efficient data storage and retrieval functions. A deep learning association analysis model is used, which uses a multi-layer neural network structure including 100 input layer nodes, 3 hidden layers (the number of nodes in each hidden layer is 50, 30, and 20, respectively), and 10 output layer nodes. The input layer receives multi-source data, and after complex operations of multiple hidden layers, the hidden layer operation uses the ReLU activation function, and the output layer outputs the results of the association relationship between data, automatically mining the complex nonlinear relationship between multi-source data. -S202: Dynamic clustering analysis - Dynamic clustering evolution algorithm is used, based on the K-Means clustering algorithm, combined with the idea of ​​genetic algorithm. Initially, five cluster centers are set. In each iteration, the distance from each parent-child data point to each cluster center is calculated, and the data points are divided into corresponding clusters according to the principle of the shortest distance. Then, the cluster centers are optimized using a genetic algorithm, and the dynamic update and optimization of the cluster centers are achieved through selection, crossover, and mutation operations. Based on the real-time updated sports data characteristics of parents and children, including sports ability (obtained through comprehensive evaluation of sports endurance, strength, and coordination indicators), sports preferences (such as the degree of preference for different types of sports such as running, ball games, and parent-child games), and emotional interaction patterns (obtained based on the analysis of parent-child communication frequency and emotional expression intensity in social interaction data), the clustering results are dynamically adjusted to accurately divide parents and children into different group categories. S3: Intervention strategy generation: S301: Customized exercise plans based on VR / AR: Based on the results of the exercise ability assessment and the exercise goals set by parents and children, we use a professional virtual reality (VR) / augmented reality (AR) technology development team to customize exclusive exercise plans for different groups of parents and children. During the development process, we fully consider the differences in parent-child exercise ability (such as dividing exercise ability into three levels: low, medium, and high, and designing exercise sessions of different difficulty and intensity for different levels), goal setting (such as improving physical fitness, strengthening parent-child relationships, and cultivating interest in sports), and adaptability to sports scenes (indoor scenes focus on space utilization and safety, and outdoor scenes focus on integration with the natural environment). For endurance-oriented parent-child groups, a virtual marathon track running challenge is designed, using AR technology to overlay virtual track instructions (track instructions are presented in clear and eye-catching graphics and text, and the distance prompt accuracy can reach ±1 meter) and competitors in real scenes (the virtual images of competitors are customized according to the challenge difficulty and goals set by parents and children, and have different running speeds and endurance performances); for strength-oriented parent-child groups, a VR-based strength training game is developed to simulate a weightlifting competition scene, with 5 weight levels, 3 different shapes of barbells, as well as realistic competition environment sound effects and audience interaction elements in the virtual scene. S302: Construction of emotional intelligent feedback system: Construct an emotional intelligent feedback system, which has the function modules of emotional dictionary construction, semantic analysis and emotional classification. The emotional dictionary collects more than 5,000 emotional words related to parent-child sports, and continuously expands and optimizes them in combination with the emotional expression data in actual sports scenes. The semantic analysis module uses a natural language processing model based on the Transformer architecture to perform grammatical and semantic analysis on the voice communication text during parent-child sports and extract key information. The emotional classification module uses deep learning algorithms, such as convolutional neural networks (CNN), to classify the facial expressions and voice emotions of parents and children, and can accurately identify 8 emotional states. The system uses natural language processing and emotion recognition technology to monitor the emotional information conveyed by the voice and expressions of parents and children during sports in real time. When the system detects changes in emotional needs, it automatically adjusts the intervention strategy. For example, when it detects that the child is frustrated and has difficulty with the difficulty of the exercise, it pushes suggestions for reducing the difficulty and encouraging speech templates to parents through a pre-set rule library. The encouraging speech templates are personalized according to the child's age and personality characteristics. S4: Dynamic adjustment of strategies: S401: Adaptive learning adjustment engine deployment: Deploy an adaptive learning adjustment engine, which has built-in time series analysis algorithms (using ARIMA (2,1,2) model) and machine learning classification algorithms (such as support vector machine SVM). Through real-time monitoring of parent-child exercise data, the data changes are analyzed at a frequency of 5 times per minute. When it is found that the athletic ability is improved (such as endurance improvement is manifested by an increase of more than 10% in exercise time for 7 consecutive days, and an increase in speed of more than 5%), or emotional needs change (such as a decrease of more than 20% in the child's positive emotional expression for a certain type of exercise in 5 consecutive exercises), the strategy adjustment mechanism is automatically triggered. If the parent-child exercise endurance is monitored to improve for two consecutive weeks, the engine will automatically increase the exercise duration and intensity based on the pre-set exercise intensity adjustment model. The exercise duration will increase by 15% of the current duration, and the intensity will be reasonably adjusted according to the professional sports training standards based on the type of exercise and the physical condition of the parent-child. If the parent-child's interest in outdoor parent-child sports decreases, the engine will recommend indoor parent-child sports programs and adjust the exercise plan based on historical data and preferences, by analyzing the user's exercise records in the application and searching for preference data. The recommended indoor parent-child sports programs include at least 3 programs that are in line with the parent-child's interests and athletic abilities. S402: Construction of community interaction and strategy sharing platform - Construction of community interaction and strategy sharing platform, which has user identity authentication (using multiple authentication methods based on mobile phone number, verification code and facial recognition), data security encryption (using AES-256 encryption algorithm to encrypt, store and transmit user data), and personalized recommendation algorithm interface functions. Using big data analysis technology, the sports experience shared by parent-child families on the platform (presented in various forms such as text, pictures, and videos) and the feedback on the effect of implementing intervention strategies (collected through questionnaires and user evaluations) are mined and analyzed. Based on the analysis results, other successful intervention strategies of similar parent-child families are recommended to families in need. The recommendation algorithm comprehensively considers the factors of eight dimensions, including parent-child sports ability, goals, emotional needs, and geographical location, to achieve the continuous evolution and sharing of personalized intervention strategies. At the same time, the platform optimizes the recommendation algorithm every 15 days, and continuously adjusts the algorithm parameters to improve the accuracy of recommendations by collecting user feedback and effect evaluation on the use of recommended strategies.

2. The method for generating a personalized intervention strategy driven by parent-child interaction exercise data according to claim 1, characterized in that: The motion tracking devices in the multi-source fusion sensing technology are wearable devices such as smart bracelets and sports watches. The bioelectric sensors are professional devices that fit the surface of the parent-child skin and can accurately capture muscle electrical signals. Environmental sensors can be flexibly deployed indoors and outdoors according to different sports scenarios.

3. The method for generating a personalized intervention strategy driven by parent-child interaction exercise data according to claim 1, characterized in that: The deep learning association analysis model adopts a multi-layer neural network structure. The input layer receives multi-source data, and after complex calculations in multiple hidden layers, the output layer outputs the results of the association relationship between the data; the dynamic clustering evolutionary algorithm is based on the K-Means clustering algorithm and combines the idea of ​​genetic algorithm to realize the dynamic update and optimization of the cluster center.

4. The method for generating a personalized intervention strategy driven by parent-child interaction exercise data according to claim 1, characterized in that: The exclusive exercise plan designed based on virtual reality (VR) / augmented reality (AR) technology must fully consider the differences in parent-child sports abilities, goal setting, and adaptability to sports scenes during the development process; the emotional intelligent feedback system has emotional dictionary construction, semantic analysis, and emotional classification functional modules to ensure accurate recognition and analysis of parent-child emotional information.

5. The method for generating a personalized intervention strategy driven by parent-child interaction exercise data according to claim 1, characterized in that: The adaptive learning adjustment engine has built-in multiple data change trend recognition algorithms, such as time series analysis algorithms and machine learning classification algorithms, to accurately judge the changes in sports data; the community interaction and strategy sharing platform has user identity authentication, data security encryption, and personalized recommendation algorithm interface functions to ensure data security and recommendation accuracy.

Images