Intelligent monitoring method for health state of athlete based on big data analysis of wearable device

Abstract

The invention relates to the technical field of big data analysis, and discloses an intelligent monitoring method for the health state of an athlete based on big data analysis of wearable equipment. According to the method, by collecting physiological data, motion data, sleep and recovery data and environment data of an athlete in real time, the health state of the athlete can be monitored in real time, the problems of low efficiency and subjectivity caused by manual recording and experience judgment in a traditional health monitoring method are solved, and through data preprocessing and feature extraction, the health state of the athlete can be monitored in real time. In addition, the health risk of the athlete can be automatically identified and the fatigue degree and the heart risk can be accurately evaluated by combining a health state evaluation model established by a deep neural network, and the method not only provides personalized exercise intensity adjustment suggestions and rest and rehabilitation guidance for the athlete, but also improves the accuracy of exercise intensity adjustment. And a continuously tracked health file is formed, and scientific training decision support is provided for a coach team, so that the body health and safety of athletes are effectively guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of big data analysis, and in particular to an intelligent monitoring method for the health status of athletes based on big data analysis of wearable devices. Background Art

[0002] With the rapid development of science and technology, wearable devices (such as smart watches, fitness trackers, etc.) are being used more and more widely in athlete health monitoring and management. These devices can collect athletes' physiological data, exercise data, sleep and recovery data, and exercise environment data in real time, providing a rich data foundation for athletes' health status. Modern sports science research shows that athletes' health status is closely related to their physiological indicators, exercise load, recovery status and external environment. These factors jointly affect athletes' performance and training results.

[0003] Traditional athlete health monitoring methods rely heavily on manual records and empirical judgments, making it difficult to achieve real-time, objective, and systematic evaluations. This approach is not only inefficient but also easily affected by subjective factors, making it impossible to detect potential health risks in athletes in a timely manner. Summary of the Invention

[0004] (1) Technical problems solved

[0005] In response to the shortcomings of the existing technology, the present invention provides an intelligent monitoring method for the health status of athletes based on big data analysis of wearable devices. By real-time collection of athletes' physiological data, exercise data, sleep and recovery data, and environmental data, it can realize real-time monitoring of athletes' health status, overcoming the inefficiency and subjectivity caused by manual recording and experience judgment in traditional health monitoring methods. Through data preprocessing and feature extraction, the high quality and consistency of the data are ensured, making the analysis results more reliable. In addition, the health status assessment model established in combination with deep neural networks can automatically identify the health risks of athletes and accurately assess fatigue levels and heart risks. This method not only provides athletes with personalized exercise intensity adjustment suggestions and rest and rehabilitation guidance, but also forms a continuously tracked health file, providing the coaching team with scientific training decision support, thereby effectively protecting the physical health and safety of athletes and improving overall sports performance.

[0006] (2) Technical solution

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for intelligently monitoring the health status of athletes based on big data analysis of wearable devices, comprising the following steps:

[0008] S1. Use wearable devices to obtain athletes' physiological data, exercise data, sleep and recovery data, and exercise environment data;

[0009] S2. Perform data preprocessing on the acquired athlete's physiological data, exercise data, sleep and recovery data, and exercise environment data. The data preprocessing includes noise filtering and missing value filling, and data consistency processing for different source data.

[0010] S3. Calculating characteristic indicators from the pre-processed athlete's physiological data, exercise data, sleep and recovery data, and environmental data, including heart rate variability, exercise intensity index, fatigue index, sleep quality index, and physical stress index, and screening the characteristic indicators using principal component analysis to obtain key characteristic data of the athlete for subsequent analysis;

[0011] S4. Use deep neural networks to train key feature data of athletes, establish an athlete health status assessment model, and use cross-validation to optimize the parameters of the athlete health status assessment model;

[0012] S5. Inputting the pre-processed and feature-extracted physiological data, exercise data, sleep and recovery data, and environmental data of athletes collected in real time into the athlete health status assessment model after parameter optimization to perform real-time prediction and dynamic change monitoring of the athlete's health status, and assess the athlete's fatigue level and cardiac risk;

[0013] S6. Based on the assessment results of the athlete's health status using the athlete's health status assessment model, provide athletes and sports instructors with suggestions on exercise intensity adjustment and rest and rehabilitation, store the athlete's health files, and form a continuously tracked athlete health monitoring data record.

[0014] Preferably, the noise filtering is achieved by the following formula:

[0015]

[0016] In the formula, Represents the value of the i-th data point after filtering, N represents the number of elements in the window, M represents the size of the filter window, x k represents the kth data point.

[0017] Preferably, the missing value filling is achieved by the following formula:

[0018]

[0019] In the formula, x i Indicates the missing position value after filling, x j represents the nearest known value to the left of the missing point, x k represents the nearest known value to the right of the missing point, t j Represents x j The corresponding time point, tk Represents x k The corresponding time point, t i Indicates the time point that needs to be filled.

[0020] Preferably, the data consistency processing for different source data is achieved by the following formula:

[0021]

[0022] In the formula, represents the normalized data, x i Indicates the missing position value after filling, x min Indicates the minimum value of the sample data that needs to be normalized, x max Indicates the maximum value of the sample data that needs to be normalized.

[0023] Preferably, the calculation formula of the heart rate variability is as follows:

[0024]

[0025] In the formula, HRV represents heart rate variability, RR i represents the time difference between the i-th heartbeats, represents the average value of the time difference between heartbeats, K represents the number of time differences between heartbeats, and i represents the index subscript.

[0026] Preferably, the calculation formula of the exercise intensity index is as follows:

[0027]

[0028] In the formula, EI represents the exercise intensity index, T represents the sampling duration, and P i Indicates the acceleration modulus value of the i-th sampling point, where i represents the index subscript.

[0029] Preferably, the calculation formula of the fatigue index is as follows:

[0030]

[0031] In the formula, EI stands for fatigue index, HRV stands for heart rate variability, and EI stands for exercise intensity index.

[0032] Preferably, the calculation formula of the sleep quality index is as follows:

[0033]

[0034] In the formula, SQI represents sleep quality index, sp represents deep sleep time, zp represents total sleep time, C represents sleep continuity, D represents the number of sleep interruptions, w1, w2, and w3 represent weight coefficients, which are dynamically assigned by the intelligent analysis chip in the wearable device.

[0035] Preferably, the calculation formula of the body pressure index is as follows:

[0036] PSI=a*HR+b*RC+c*FI

[0037] In the formula, PSI represents the body stress index, HR represents the heart rate, RC represents the respiratory rate, FI represents the fatigue index, a, b, and c represent the weight coefficients corresponding to the calculation indicators, and are assigned through big data analysis.

[0038] Preferably, the athlete health status assessment model is as follows:

[0039]

[0040] In the formula, Represents the model prediction output result, X represents the input feature vector, θ represents the parameters obtained by model training, and DeepNN represents the deep neural network model.

[0041] Compared with the existing technology, the present invention provides an intelligent monitoring method for athletes' health status based on big data analysis of wearable devices, which has the following beneficial effects:

[0042] This invention can achieve real-time monitoring of athletes' health status by collecting athletes' physiological data, exercise data, sleep and recovery data, and environmental data in real time, overcoming the inefficiency and subjectivity caused by manual recording and experience judgment in traditional health monitoring methods. Through data preprocessing and feature extraction, it ensures the high quality and consistency of data, making the analysis results more reliable. In addition, the health status assessment model established in combination with deep neural networks can automatically identify athletes' health risks and accurately assess fatigue levels and heart risks. This method not only provides athletes with personalized exercise intensity adjustment suggestions and rest and rehabilitation guidance, but also forms a continuously tracked health file, providing the coaching team with scientific training decision support, thereby effectively protecting the physical health and safety of athletes and improving overall sports performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 Schematic diagram of the steps of the method of the present invention. DETAILED DESCRIPTION

[0044] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0045] Traditional athlete health monitoring methods rely heavily on manual recording and empirical judgment, making it difficult to achieve real-time, objective, and systematic evaluation. This approach is not only inefficient but also easily influenced by subjective factors, making it difficult to promptly identify potential health risks faced by athletes. Therefore, we propose an intelligent athlete health status monitoring method based on big data analysis using wearable devices. Figure 1 , the method comprises the following steps:

[0046] S1. Use wearable devices to obtain athletes' physiological data, exercise data, sleep and recovery data, and exercise environment data;

[0047] By integrating multimodal sensing technology and advanced data acquisition systems, the multifunctional sensor array in wearable devices is used to conduct continuous, real-time, and multi-dimensional data monitoring of athletes. Specifically, high-precision heart rate sensors, photoplethysmography sensors (PPG) or electrocardiogram (ECG) modules are used to collect athletes' heart rate, heart rate variability and other cardiovascular parameters in real time. Motion sensors such as three-axis accelerometers, gyroscopes, and magnetometers are equipped to accurately record sports performance indicators such as motion trajectory, exercise intensity, acceleration, angular velocity and exercise frequency. A sleep monitoring module is introduced to achieve deep sleep, light sleep, and rapid eye movement (REM) sleep through sleep monitoring sensors, blood oxygen saturation (SpO2) sensors, and exercise / sleep cycle monitoring algorithms. Accurate analysis of sleep structure such as sleep, while deploying environmental sensors to detect air quality (PM2.5, PM10, VOC, CO2, etc.), temperature and humidity, air pressure and ultraviolet intensity, providing data support for sports environment safety and sports performance. All collected physiological, sports, sleep and environmental data are processed in real time by an embedded microprocessor and continuously transmitted to the background data management platform using wireless communication technologies such as Bluetooth or Wi-Fi. The multi-channel data synchronization mechanism ensures the consistency of information timing and realizes efficient, continuous and comprehensive data collection. The core of this technical means lies in multi-source, multi-modal sensor fusion technology and efficient data collection architecture, providing a solid data foundation for athlete health status assessment, sports performance optimization and personalized training programs;

[0048] S2. Perform data preprocessing on the acquired athlete's physiological data, exercise data, sleep and recovery data, and exercise environment data. The data preprocessing includes noise filtering and missing value filling, and data consistency processing for different source data.

[0049] The multi-source and multi-modal data collected by athletes, including physiological parameters, exercise status, sleep and recovery indicators, and environmental information, are systematically pre-processed to ensure the accuracy and reliability of subsequent analysis. In terms of noise filtering, a variety of signal pre-processing algorithms are used, such as the moving average filter, which can be expressed as By smoothing the adjacent data points, high-frequency noise and sharp anomalies are suppressed, while the median filter selects the median value by sorting the data in the window to effectively suppress impulse noise. For missing value filling, linear interpolation is often used. According to the missing point and its nearest known adjacent point, the formula is used. Interpolation obtains missing data to ensure the continuity of the time series. In addition, the missing values ​​can be filled by mean value or interpolation to ensure data integrity. In the process of integrating different source data, in order to ensure the consistency and comparability of the data, normalization is first performed, such as the maximum and minimum normalization formula Mapping the features of each data source to a unified range to avoid the impact of scale differences on analysis results. The core of this set of preprocessing measures is to effectively filter noise, supplement missing information, and achieve standardization and consistency of multi-source data, thereby providing stable and reliable basic data for subsequent feature extraction and model building.

[0050] S3. Calculating characteristic indicators from the pre-processed athlete's physiological data, exercise data, sleep and recovery data, and environmental data, including heart rate variability, exercise intensity index, fatigue index, sleep quality index, and physical stress index, and screening the characteristic indicators using principal component analysis to obtain key characteristic data of the athlete for subsequent analysis;

[0051] In the feature extraction stage of the pre-processed multi-source athlete data, a variety of scientific calculation methods and statistical analysis techniques are used to extract representative and discriminative feature indicators from it, so as to more accurately reflect the health status and sports performance of the athletes. Specifically, the heart rate variability is first calculated. Common indicators include the standard deviation of heart rate variability, and its formula is This indicator can effectively reflect the regulatory ability of the autonomic nervous system and is an important physiological parameter for evaluating stress and fatigue. Then, by analyzing the acceleration information in the motion data, the exercise intensity index is calculated, such as the average acceleration value. It can measure the intensity of exercise load and assist in evaluating the effect of exercise training. The fatigue index is calculated by combining heart rate variability and exercise intensity. The formula is: This indicator can highlight the state of fatigue. The higher the value, the greater the risk of fatigue and overtraining of the athlete. In terms of sleep and recovery, the sleep quality index is calculated using a weighted average method, combining the proportion of deep sleep time, sleep continuity and the number of interruptions, using the formula:

[0052]

[0053] Realize multi-dimensional sleep quality assessment, weight w i The setting can highlight the importance of different indicators. During this period, a physical stress index is established, which can be calculated based on a combination of heart rate, respiratory rate and fatigue indicators. The formula is PSI = a*HR+b*RC+c*FI, which is used to comprehensively reflect the stress state of athletes during training or competition. After these multivariate feature variables are extracted, the data dimension is huge. In order to avoid redundant information interfering with subsequent model learning, principal component analysis (PCA) is used to reduce the dimension and screen the feature indicators. Its core formula is Z = XW, where X is the original feature matrix and W is the feature load matrix. The principal component that best represents the original information is found, thereby extracting the key features with the greatest representativeness and distinguishing ability. After using PCA screening, it not only effectively reduces data noise and redundancy, but also enhances the stability and predictive performance of subsequent models, which is of great significance to improving analysis efficiency and model generalization ability. In addition, the screened key features can also help researchers to have a deeper understanding of the internal laws of athletes' status and realize personalized training and precise health management.

[0054] S4. Use deep neural networks to train key feature data of athletes, establish an athlete health status assessment model, and use cross-validation to optimize the parameters of the athlete health status assessment model;

[0055] In the process of building the athlete health status assessment model, the deep neural network (DNN) is used as the core technology platform. Its complex nonlinear mapping capability is used to perform end-to-end training on the key indicator data of athletes that have been feature-screened and preprocessed, thereby establishing an accurate and reliable health status prediction model. The expression is By designing a multi-layer hidden layer network structure and combining activation functions such as ReLU, LeakyReLU, Sigmoid, etc., the expressive power of the model is enhanced, regularization techniques (such as Dropout and L2 regularization) are introduced to reduce the risk of overfitting, and the back propagation algorithm is combined with advanced optimizers (such as Adam, RMSProp) to dynamically adjust the network parameters to improve the training efficiency while ensuring the output of the model. In order to further improve the generalization ability and parameter stability of the model and effectively avoid overfitting in different data partitions, the cross-validation technology is used to divide the data into multiple folds, which are trained and evaluated in turn as training sets and validation sets. The model is optimized through multiple rounds of validation. Hyperparameters (such as learning rate, network depth, and number of neurons per layer) are adjusted and the optimal parameter combination is selected. The biggest advantage of using deep neural networks to predict athlete health status lies in their powerful automatic feature extraction capabilities, which can identify complex nonlinear relationships in data and are more expressive and accurate than traditional linear models. At the same time, the multi-layer structure helps capture potential abstract features and achieve a deep understanding of athletes' physiological, sports, and environmental information, thereby providing scientific and reliable athlete health status assessments and providing strong technical support for sports physiological management, personalized training, and risk warnings. The implementation of this solution will greatly enhance the intelligence level of health monitoring and the accuracy of predictions, and promote the deep integration of sports medicine and artificial intelligence.

[0056] S5. Inputting the pre-processed and feature-extracted physiological data, exercise data, sleep and recovery data, and environmental data of athletes collected in real time into the athlete health status assessment model after parameter optimization to perform real-time prediction and dynamic change monitoring of the athlete's health status, and assess the athlete's fatigue level and cardiac risk;

[0057] The athletes' physiological data, exercise data, sleep and recovery data, and environmental data collected in real time are standardized to adapt to the model input format, and then input into the model in batch or streaming mode. The evaluation model with optimized parameters is used for analysis to obtain the athlete's current health status assessment results. The system tracks the changes in prediction indicators in real time, thereby dynamically monitoring the athlete's fatigue level, cardiac risk level and other key health indicators. Potential health risks are inferred by combining thresholds or risk models, making it easier for athletes and coaches to develop personalized rest, training and intervention plans. This technical framework integrates real-time data collection, scientific feature engineering, deep learning models and multi-indicator risk assessment to ensure the timeliness and accuracy of monitoring, provide intelligent, scientific and dynamic decision-making support for athlete health management, effectively prevent sudden sports-related risks, and improve overall training performance and health protection levels.

[0058] S6. Based on the athlete health status assessment results of the athlete health status assessment model, provide athletes and sports instructors with exercise intensity adjustment suggestions and rest and recovery suggestions, and store athlete health files to form a continuous tracking athlete health monitoring data record;

[0059] Based on the output results of the athlete health status assessment model, the system uses intelligent decision-making algorithms (such as rule engines, machine learning classifiers or multi-objective optimization models) to provide athletes and sports instructors with personalized exercise intensity adjustment suggestions (such as adding or subtracting training loads, adjusting exercise frequency and intensity) and scientific rest and rehabilitation plans (including rest time, rehabilitation exercise recommendations, nutritional guidance, etc.) to achieve reasonable training plans and health management. This process combines the real-time status, historical data and individual characteristics of athletes, and uses expert knowledge bases or machine learning models to ensure that the suggestions are scientific, effective and operational. At the same time, in order to achieve continuous health status tracking, all athletes' health assessment results, training plan adjustment records, rehabilitation measures and related indicators are stored in dynamically updated athlete health files, using high An efficient database management system (such as a relational database or a time-series database) is used to achieve standardized storage and security protection of data. Combined with data warehouse and big data analysis technology, the system continuously accumulates historical data on athlete health monitoring to form a complete athlete health file, providing data support for subsequent trend analysis, personalized management and early warning. Through visual dashboards and mobile applications, athletes and coaches can easily view and adjust training plans. At the same time, the system can automatically optimize the recommendation model based on accumulated data, gradually improving the scientificity and practicality of personalized plans. The integration of this series of technical means has achieved dynamic management, scientific intervention and full-cycle tracking of athletes' health status, ensuring the safety and maximization of sports training, and providing solid technical support for the long-term development and physical health of athletes.

[0060] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent monitoring method for athlete health status based on big data analysis of wearable devices, characterized in that: The following steps are involved: S1. Use wearable devices to obtain athletes' physiological data, exercise data, sleep and recovery data, and exercise environment data; S2. Perform data preprocessing on the acquired athlete's physiological data, exercise data, sleep and recovery data, and exercise environment data. The data preprocessing includes noise filtering and missing value filling, and data consistency processing for different source data. S3. Calculating characteristic indicators from the pre-processed athlete's physiological data, exercise data, sleep and recovery data, and environmental data, including heart rate variability, exercise intensity index, fatigue index, sleep quality index, and physical stress index, and screening the characteristic indicators using principal component analysis to obtain key characteristic data of the athlete for subsequent analysis; S4. Use deep neural networks to train key feature data of athletes, establish an athlete health status assessment model, and use cross-validation to optimize the parameters of the athlete health status assessment model; S5. Inputting the pre-processed and feature-extracted physiological data, exercise data, sleep and recovery data, and environmental data of athletes collected in real time into the athlete health status assessment model after parameter optimization to perform real-time prediction and dynamic change monitoring of the athlete's health status, and assess the athlete's fatigue level and cardiac risk; S6. Based on the assessment results of the athlete's health status using the athlete's health status assessment model, provide athletes and sports instructors with suggestions on exercise intensity adjustment and rest and rehabilitation, store the athlete's health files, and form a continuously tracked athlete health monitoring data record.

2. The method for intelligently monitoring athlete health status based on wearable device big data analysis according to claim 1, characterized in that: The noise filtering is achieved by the following formula: In the formula, Represents the value of the i-th data point after filtering, N represents the number of elements in the window, M represents the size of the filter window, x k represents the kth data point.

3. The method for intelligently monitoring athlete health status based on wearable device big data analysis according to claim 2, characterized in that: The missing value filling is achieved by the following formula: In the formula, x i Indicates the missing position value after filling, x j represents the nearest known value to the left of the missing point, x k represents the nearest known value to the right of the missing point, t j Represents x j The corresponding time point, t k Represents x k The corresponding time point, t i Indicates the time point that needs to be filled.

4. The method for intelligently monitoring athlete health status based on wearable device big data analysis according to claim 3, characterized in that: The data consistency processing of different source data is achieved by the following formula: In the formula, represents the normalized data, x i Indicates the missing position value after filling, x min Indicates the minimum value of the sample data that needs to be normalized, x max Indicates the maximum value of the sample data that needs to be normalized.

5. The method for intelligently monitoring athlete health status based on wearable device big data analysis according to claim 4, characterized in that: The calculation formula of the heart rate variability is as follows: In the formula, HRV represents heart rate variability, RR i represents the time difference between the i-th heartbeats, represents the average value of the time difference between heartbeats, K represents the number of time differences between heartbeats, and i represents the index subscript.

6. The method for intelligently monitoring athlete health status based on wearable device big data analysis according to claim 5, characterized in that: The calculation formula of the exercise intensity index is as follows: In the formula, EI represents the exercise intensity index, T represents the sampling duration, and P i Indicates the acceleration modulus value of the i-th sampling point, where i represents the index subscript.

7. The method for intelligently monitoring athlete health status based on wearable device big data analysis according to claim 6, characterized in that: The calculation formula of the fatigue index is as follows: In the formula, FI stands for fatigue index, HRV stands for heart rate variability, and EI stands for exercise intensity index.

8. The method for intelligently monitoring athlete health status based on wearable device big data analysis according to claim 7, characterized in that: The calculation formula of the sleep quality index is as follows: In the formula, SQI represents sleep quality index, sp represents deep sleep time, zp represents total sleep time, C represents sleep continuity, D represents the number of sleep interruptions, w1, w2, and w3 represent weight coefficients, which are dynamically assigned by the intelligent analysis chip in the wearable device.

9. The method for intelligently monitoring athlete health status based on wearable device big data analysis according to claim 8, characterized in that: The calculation formula of the body pressure index is as follows: PSI=a*HR+b*RC+c*FI In the formula, PSI represents the body stress index, HR represents the heart rate, RC represents the respiratory rate, FI represents the fatigue index, a, b, and c represent the weight coefficients corresponding to the calculation indicators, and are assigned through big data analysis.

10. The method for intelligently monitoring athlete health status based on wearable device big data analysis according to claim 9, characterized in that: The athlete health status assessment model is as follows: In the formula, Represents the model prediction output result, X represents the input feature vector, θ represents the parameters obtained by model training, and DeepNN represents the deep neural network model.

Images