Sports health monitoring method and system of smart watch

Abstract

The invention discloses an exercise health monitoring method and system of a smart watch, and relates to the technical field of smart watches, and the method comprises the steps: carrying out the time synchronization of continuous data, and enabling all monitoring indexes to be aligned on the same time axis; carrying out real-time analysis and anomaly detection on the heart rate and blood oxygen saturation data by utilizing a machine learning algorithm; the heart rate variability index of the user is evaluated, the blood oxygen saturation data and the body surface temperature data are integrated, and a pressure evaluation model is constructed; in combination with height and weight information input by the user, constructing a deep learning model to identify a motion mode of the user; the heart rate and blood oxygen health indexes of the user are tracked and analyzed by utilizing a time sequence analysis technology, the future health trend is predicted, and potential health problems are explored. By arranging the data prediction module, the heart rate, blood oxygen and other health indexes of the user can be tracked and analyzed, the future health trend is predicted, and potential health problems are explored.

Description

Technical Field

[0001] The present invention relates to the technical field of smart watches, and in particular to a sports health monitoring method and system for a smart watch. Background Art

[0002] With the rapid development of science and technology, smart wearable devices have gradually been integrated into people's daily lives. Among them, smart watches, as a representative of fashion, convenience and functionality, have received widespread attention and love. With the increasing awareness of health, smart watches are not only a tool to check time, but also an important assistant for people to monitor their physical health and manage their sports life.

[0003] Smart watches provide a lot of health data, but it is not easy for ordinary users to correctly interpret this data. Users may lack professional medical knowledge and cannot accurately understand the meaning of the data and potential health risks. Users have different health needs and different requirements for the functions and performance of smart watches. Summary of the invention

[0004] In order to solve the above technical problems, a sports health monitoring method and system for a smart watch are provided. This technical solution solves the problems raised in the above background technology.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is: A sports health monitoring method for a smart watch, comprising: The heart rate sensor and blood oxygen sensor monitor the user's heart rate and blood oxygen saturation in real time. The accelerometer and gyroscope record the user's movement trajectory, steps, speed and distance information. The microphone detects environmental noise. The temperature sensor monitors the body surface temperature. Time synchronization of continuous data to align each monitoring indicator on the same time axis; Use machine learning algorithms to perform real-time analysis and anomaly detection on heart rate and blood oxygen saturation data to identify abnormal fluctuations; Based on the heart rate sensor data, the user's heart rate variability index is evaluated, and the blood oxygen saturation data and body surface temperature data are combined to build a stress assessment model and output stress management suggestions; Based on the accelerometer and gyroscope data, combined with the height and weight information entered by the user, a deep learning model is built to identify the user's movement pattern and calculate the calories consumed; Provide personalized exercise plans and energy consumption analysis reports by analyzing the user's historical exercise habits and physical fitness level; Using time series analysis technology, track and analyze the user's heart rate and blood oxygen health indicators, predict future health trends, and discover potential health problems; Upload various health data to the cloud server for storage and backup.

[0006] Preferably, the real-time analysis and abnormality detection of heart rate and blood oxygen saturation data using a machine learning algorithm to identify abnormal fluctuations therein specifically includes: The heart rate and blood oxygen saturation data were cleaned to remove values ​​beyond the physiological range. For missing values, the front and back value filling method was used to complete them. Standardize the data, convert it to the same order of magnitude, convert the data to a mean of 0 and a standard deviation in the range of [0,1]; Extracting derived features of the heart rate and blood oxygen saturation data, wherein the derived features include a rate of change of the heart rate and a fluctuation range of the blood oxygen saturation; Based on the characteristics of heart rate and blood oxygen saturation data, the long short-term memory network algorithm was selected as the machine learning algorithm for model training; Build a training dataset containing normal and abnormal heart rate and blood oxygen saturation data. Normal data comes from daily monitoring data of healthy users, and abnormal data comes from data of users with known health problems. The selected machine learning algorithm is trained using the training data set to obtain a model for detecting abnormal fluctuations in the user's heart rate and blood oxygen saturation data; The smart watch collects heart rate and blood oxygen saturation data in real time to form a data stream, extracts features from the data stream, and obtains a real-time feature vector; The real-time feature vector is input into the abnormal fluctuation detection model of the user's heart rate and blood oxygen saturation data to obtain a prediction result, which is the probability of normal or abnormality of the user's current physiological state; Based on the prediction results, abnormal fluctuations in heart rate and blood oxygen saturation are identified and abnormal alarms are triggered. The abnormal alarms remind users through watch vibrations, screen displays, and mobile phone app notifications. Based on the accumulation of user data and the discovery of new physiological characteristics, the machine learning model is updated to adapt to changes in user physiological status and new monitoring needs.

[0007] Preferably, the use of time series analysis technology to track and analyze the user's heart rate and blood oxygen health indicators, predict future health trends, and discover potential health problems specifically includes: Conduct visual analysis of the data, draw time series graphs, autocorrelation graphs, and partial autocorrelation graphs, and preliminarily determine the stability and trend of the data; The order of the autoregressive term is determined by observing the partial autocorrelation graph. The partial autocorrelation graph decays and approaches 0 after lagging the order. The optimal order value is selected by comparing the model fitting effects under different order values. The order of the moving average term is determined by observing the autocorrelation graph. The autocorrelation graph decays and approaches 0 after lagging the order. The optimal order value is selected by comparing the model fitting effects under different order values. Perform a stationarity test to determine whether the data is not stationary. If so, perform a difference process to convert it into a stationary sequence. The number of differences is determined by at least one difference and stationarity test. If not, no output is made. Based on the order of the autoregressive term, the number of differences and the order of the moving average term, an autoregressive integrated moving average model is constructed; Determine whether there is correlation between the residual sequences of the autoregressive integrated moving average model. If so, adjust the order of the model's autoregressive term, the number of differences, and the order of the moving average term. If not, no output is made. Using the trained model, the user's future heart rate and blood oxygen saturation are predicted in the short term, so that the user can understand his or her health status and take corresponding preventive measures; By analyzing long-term trends and cyclical changes in historical data, we can make long-term predictions about users' future health conditions, plan health management strategies for users, and prevent potential health problems.

[0008] Furthermore, a sports health monitoring system for a smart watch is proposed, which is used to implement the sports health monitoring method for the smart watch as described above, comprising: A data acquisition module, which is used to monitor the user's heart rate and blood oxygen saturation in real time through a heart rate sensor and a blood oxygen sensor, record the user's movement trajectory, steps, speed and distance information through an accelerometer and a gyroscope, detect environmental noise using a microphone, and monitor body surface temperature through a temperature sensor; A data processing module, which is used to synchronize the continuous data, align the monitoring indicators on the same time axis, and use machine learning algorithms to perform real-time analysis and anomaly detection on the heart rate and blood oxygen saturation data to identify abnormal fluctuations therein; Health monitoring module, which is used to evaluate the user's heart rate variability index based on heart rate sensor data, and to build a stress assessment model based on blood oxygen saturation data and body surface temperature data, output stress management suggestions, build a deep learning model based on accelerometer and gyroscope data, combined with the height and weight information input by the user to identify the user's exercise pattern, and calculate the calories consumed, and provide personalized exercise plans and energy consumption analysis reports by analyzing the user's historical exercise habits and physical fitness level; The data prediction module is used to use time series analysis technology to track and analyze the user's heart rate and blood oxygen health indicators, predict future health trends, discover potential health problems, and upload various health data to the cloud server for storage and backup.

[0009] Compared with the prior art, the present invention has the following beneficial effects: Time synchronization of continuous data aligns each monitoring indicator on the same timeline, ensuring data consistency and accuracy. This helps to more accurately analyze the correlation and changing trends among indicators. The user's heart rate variability indicator is evaluated based on heart rate sensor data, and a stress assessment model is constructed based on blood oxygen saturation data and body surface temperature data. This function enables users to understand their stress status and obtain corresponding stress management suggestions, which helps to reduce stress and maintain physical and mental health. Time series analysis technology is used to track and analyze users' heart rate, blood oxygen and other health indicators, predict future health trends, and enable users to understand their health status in advance, take corresponding preventive measures, and avoid potential health problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A flow chart of the sports health monitoring method of the smart watch of the present invention; Figure 2 A flow chart of a method for time synchronization of continuous data according to the present invention; Figure 3 This is a flow chart of the method for real-time analysis and abnormality detection of heart rate and blood oxygen saturation data of the present invention; Figure 4 A flow chart of a method for evaluating a user's heart rate variability index and building a stress evaluation model by integrating blood oxygen saturation data and body surface temperature data according to the present invention; Figure 5 A flow chart of the method for constructing a deep learning model to identify a user's motion pattern of the present invention; Figure 6 This is a flow chart of the method for tracking and analyzing the user's heart rate and blood oxygen health indicators according to the present invention. DETAILED DESCRIPTION

[0011] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0012] Reference Figure 1 As shown, a sports health monitoring method of a smart watch includes: The heart rate sensor and blood oxygen sensor monitor the user's heart rate and blood oxygen saturation in real time. The accelerometer and gyroscope record the user's movement trajectory, steps, speed and distance information. The microphone detects environmental noise. The temperature sensor monitors the body surface temperature. Time synchronization of continuous data to align each monitoring indicator on the same time axis; Use machine learning algorithms to perform real-time analysis and anomaly detection on heart rate and blood oxygen saturation data to identify abnormal fluctuations; Based on the heart rate sensor data, the user's heart rate variability index is evaluated, and the blood oxygen saturation data and body surface temperature data are combined to build a stress assessment model and output stress management suggestions; Based on the accelerometer and gyroscope data, combined with the height and weight information entered by the user, a deep learning model is built to identify the user's movement pattern and calculate the calories consumed; Provide personalized exercise plans and energy consumption analysis reports by analyzing the user's historical exercise habits and physical fitness level; Using time series analysis technology, track and analyze the user's heart rate and blood oxygen health indicators, predict future health trends, and discover potential health problems; Upload various health data to the cloud server for storage and backup.

[0013] Reference Figure 2 As shown in the figure, the continuous data is synchronized in time, and each monitoring indicator is aligned on the same time axis, including: When the smartwatch is first started or the user changes the device, the time is calibrated and synchronized with the Internet time server through the Network Time Protocol; The smartwatch is connected to the smartphone via Bluetooth or Wi-Fi, obtains the smartphone’s timestamp, and sets it as the smartwatch’s system time; Each sensor records an initial timestamp when it is initialized, and uses it as the time reference for subsequent data collection; When the sensor collects a set of data, a timestamp is automatically added to the set of data based on the initial timestamp when the sensor was initialized and the current sampling interval; Establish a time synchronization mechanism. The time synchronization thread checks the data buffer of each sensor and aligns the data to the same time axis according to the timestamp; For heart rate data, linear interpolation is used to estimate the heart rate value between two sampling points; For the accelerometer data, resample it to reduce the data frequency so that it is aligned with the time axis of other sensor data; Dynamically adjust the sampling frequency of sensors based on the user's activity status and monitoring needs. When the user is exercising vigorously, increase the sampling frequency of the accelerometer and gyroscope. By comparing with the network time protocol server, the accuracy of the time synchronization mechanism can be verified.

[0014] For heart rate data, the system can use linear interpolation to estimate the heart rate value between two sampling points. This method can improve the density and accuracy of heart rate data and provide more valuable information for subsequent analysis. If the frequency of accelerometer data is inconsistent with other sensor data, the system can perform resampling operations to align the frequency of accelerometer data with the time axis of other sensor data by reducing the frequency of accelerometer data. The resampling method can be nearest neighbor interpolation, bilinear interpolation, or cubic convolution interpolation.

[0015] Reference Figure 3 As shown in the figure, the machine learning algorithm is used to perform real-time analysis and anomaly detection on the heart rate and blood oxygen saturation data, and the abnormal fluctuations are identified, including: The heart rate and blood oxygen saturation data were cleaned to remove values ​​beyond the physiological range. For missing values, the front and back value filling method was used to complete them. Standardize the data, convert it to the same order of magnitude, convert the data to a mean of 0 and a standard deviation in the range of [0,1]; Extracting derived features of the heart rate and blood oxygen saturation data, wherein the derived features include a rate of change of the heart rate and a fluctuation range of the blood oxygen saturation; Based on the characteristics of heart rate and blood oxygen saturation data, the long short-term memory network algorithm was selected as the machine learning algorithm for model training; Build a training dataset containing normal and abnormal heart rate and blood oxygen saturation data. Normal data comes from daily monitoring data of healthy users, and abnormal data comes from data of users with known health problems. The selected machine learning algorithm is trained using the training data set to obtain a model for detecting abnormal fluctuations in the user's heart rate and blood oxygen saturation data; The smart watch collects heart rate and blood oxygen saturation data in real time to form a data stream, extracts features from the data stream, and obtains a real-time feature vector; The real-time feature vector is input into the abnormal fluctuation detection model of the user's heart rate and blood oxygen saturation data to obtain a prediction result, which is the probability of normal or abnormality of the user's current physiological state; Based on the prediction results, abnormal fluctuations in heart rate and blood oxygen saturation are identified and abnormal alarms are triggered. The abnormal alarms remind users through watch vibrations, screen displays, and mobile phone app notifications. Based on the accumulation of user data and the discovery of new physiological characteristics, the machine learning model is updated to adapt to changes in user physiological status and new monitoring needs.

[0016] The normal physiological range of heart rate and blood oxygen saturation was set. The normal range of heart rate was 60-100 beats / minute (for adults at rest), and the normal range of blood oxygen saturation was 95%-100%. The heart rate and blood oxygen saturation data sets were traversed to remove values ​​beyond the set range to ensure the accuracy of the data. For missing values ​​in the data set, the previous and next value filling method was used to fill in the missing values, that is, the missing value was replaced with the average of the previous and next valid values ​​to maintain the continuity of the data.

[0017] Reference Figure 4 As shown in the figure, based on the heart rate sensor data, the user's heart rate variability index is evaluated, and the blood oxygen saturation data and body surface temperature data are combined to build a stress assessment model, and the output stress management suggestions include: By obtaining the adjacent R wave intervals and N wave intervals in the heart rate data, the root mean square value of the difference between the average R wave interval and the adjacent NN interval is calculated, and the sum of the two values ​​is added to obtain the heart rate variability index; By performing Fourier transform on the heart rate variability signal and analyzing the power spectrum density of each frequency band, high frequencies are related to parasympathetic nerve activity, while low frequencies are related to sympathetic nerve activity; A heart rate variability index measurement is performed once when the user is still asleep to establish a personal baseline heart rate variability index value; Determine whether the real-time heart rate variability index value is lower than the baseline heart rate variability index value, if so, output that the user is in a high stress state, if not, no output; Determine whether the user's blood oxygen saturation is lower than a preset blood oxygen saturation threshold under high stress. If so, output that the user is in a hypoxic state. If not, no output is made. Determine whether the user's body surface temperature under high stress is higher than a preset body surface temperature threshold. If so, output that the user is in a high consumption state. If not, no output is made. Heart rate variability index, blood oxygen saturation and body surface temperature were selected as input variables of the stress assessment model, and the support vector machine method was used to build the stress assessment model to learn the relationship between the variables and the stress level; For users with high stress levels, meditation, deep breathing, and yoga are recommended to relieve stress and enhance parasympathetic nervous system activity; For users who are in hypoxic state, it is recommended that users perform moderate aerobic exercise; For users who are in a high-consumption state, it is recommended that they avoid overexertion, take rest, and adjust their breathing rhythm.

[0018] The first measurement when the user has not gotten up and is in a quiet state is recorded as the baseline heart rate variability index value. The user's heart rate variability is monitored regularly or continuously and compared with the baseline value. If the real-time value is lower than the baseline value, it is determined that the user may be in a high-stress state; if the user's blood oxygen saturation in a high-stress state is lower than a preset threshold (such as 95%), it is determined to be a hypoxic state; if the user's body surface temperature in a high-stress state is higher than a preset threshold (such as 37.5°C), it is determined to be a high-consumption state.

[0019] Reference Figure 5 As shown in the figure, based on the accelerometer and gyroscope data, combined with the height and weight information entered by the user, a deep learning model is built to identify the user's movement pattern and calculate the calories consumed, including: The features to be identified are extracted from the accelerometer and gyroscope data and used as input to the deep learning model. The features to be identified include cadence, stride length, and rotation angle.

[0020] A deep learning model for motion pattern recognition is built based on a convolutional neural network. The model structure includes convolutional layers, pooling layers, and fully connected layers. The convolutional layers are used to extract features, the pooling layers are used to reduce feature dimensions, and the fully connected layers are used for classification. Divide the preprocessed accelerometer and gyroscope data into training, validation, and test sets; The model is trained using the training set data, and the model parameters are updated through a back-propagation algorithm. The newly acquired accelerometer and gyroscope data are input into the trained model, and the model outputs the user's current motion mode, which includes walking, running, and cycling. Use the basal metabolic rate calculation formula and the exercise calorie calculation formula to calculate the total calorie consumption during exercise.

[0021] The basal metabolic rate calculation formula is: , In the formula, is the basal metabolic rate, is the user's weight, is the user's height, is the user's age; The calculation formula for the calories consumed during exercise is: , In the formula, Burn calories for exercise, The metabolic equivalent of walking is about 3.0, and the metabolic equivalent of running is about 8.0. For exercise time; The total calorie expenditure is: , In the formula, is the total calories burned during exercise.

[0022] Reference Figure 6 As shown in the figure, time series analysis technology is used to track and analyze the user's heart rate and blood oxygen health indicators, predict future health trends, and discover potential health problems, including: Conduct visual analysis of the data, draw time series graphs, autocorrelation graphs, and partial autocorrelation graphs, and preliminarily determine the stability and trend of the data; The order of the autoregressive term is determined by observing the partial autocorrelation graph. The partial autocorrelation graph decays and approaches 0 after lagging the order. The optimal order value is selected by comparing the model fitting effects under different order values. The order of the moving average term is determined by observing the autocorrelation graph. The autocorrelation graph decays and approaches 0 after lagging the order. The optimal order value is selected by comparing the model fitting effects under different order values. Perform a stationarity test to determine whether the data is not stationary. If so, perform a difference process to convert it into a stationary sequence. The number of differences is determined by at least one difference and stationarity test. If not, no output is made. Based on the order of the autoregressive term, the number of differences and the order of the moving average term, an autoregressive integrated moving average model is constructed; Determine whether there is correlation between the residual sequences of the autoregressive integrated moving average model. If so, adjust the order of the model's autoregressive term, the number of differences, and the order of the moving average term. If not, no output is made. Using the trained model, the user's future heart rate and blood oxygen saturation are predicted in the short term, so that the user can understand his or her health status and take corresponding preventive measures; By analyzing long-term trends and cyclical changes in historical data, we can make long-term predictions about users' future health conditions, plan health management strategies for users, and prevent potential health problems.

[0023] Combining the long-term trends and cyclical changes in historical data, we make long-term (such as several months to several years) predictions of users' future health conditions. Based on the long-term prediction results, we develop personalized health management strategies for users, such as exercise plans, dietary adjustments, etc., to prevent potential health problems. We regularly update data and refit the model to ensure the accuracy and timeliness of the prediction results. At the same time, we adjust the health management strategy based on user feedback and actual conditions.

[0024] Further, based on the same inventive concept as the above-mentioned sports health monitoring method of the smart watch, the present solution also proposes a sports health monitoring system for a smart watch, comprising: A data acquisition module, which is used to monitor the user's heart rate and blood oxygen saturation in real time through a heart rate sensor and a blood oxygen sensor, record the user's movement trajectory, steps, speed and distance information through an accelerometer and a gyroscope, detect environmental noise using a microphone, and monitor body surface temperature through a temperature sensor; A data processing module, which is used to synchronize the continuous data, align the monitoring indicators on the same time axis, and use machine learning algorithms to perform real-time analysis and anomaly detection on the heart rate and blood oxygen saturation data to identify abnormal fluctuations therein; Health monitoring module, which is used to evaluate the user's heart rate variability index based on heart rate sensor data, and to build a stress assessment model based on blood oxygen saturation data and body surface temperature data, output stress management suggestions, build a deep learning model based on accelerometer and gyroscope data, combined with the height and weight information input by the user to identify the user's exercise pattern, and calculate the calories consumed, and provide personalized exercise plans and energy consumption analysis reports by analyzing the user's historical exercise habits and physical fitness level; The data prediction module is used to use time series analysis technology to track and analyze the user's heart rate and blood oxygen health indicators, predict future health trends, discover potential health problems, and upload various health data to the cloud server for storage and backup.

[0025] Furthermore, the present solution also proposes a computer-readable storage medium on which a computer-readable program is stored. When the computer-readable program is called, the above-mentioned sports health monitoring method of the smart watch is executed.

[0026] It is understandable that the storage medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid state drive (SSD).

[0027] To sum up, the advantages of the present invention are: time synchronization of continuous data, alignment of various monitoring indicators on the same time axis, ensuring the consistency and accuracy of the data, which helps to more accurately analyze the correlation and change trend between the indicators, evaluate the user's heart rate variability indicators based on the heart rate sensor data, and integrate the blood oxygen saturation data and body surface temperature data to build a stress assessment model. This function enables users to understand their own stress status and obtain corresponding stress management suggestions, which helps to reduce stress and maintain physical and mental health. It uses time series analysis technology to track and analyze the user's heart rate, blood oxygen and other health indicators, predict future health trends, and enable users to understand their health status in advance, take corresponding preventive measures, and avoid the occurrence of potential health problems.

[0028] 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 to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A sports health monitoring method for a smart watch, characterized in that: include: The heart rate sensor and blood oxygen sensor monitor the user's heart rate and blood oxygen saturation in real time. The accelerometer and gyroscope record the user's movement trajectory, steps, speed and distance information. The microphone detects environmental noise. The temperature sensor monitors the body surface temperature. Time synchronization of continuous data to align each monitoring indicator on the same time axis; Use machine learning algorithms to perform real-time analysis and anomaly detection on heart rate and blood oxygen saturation data to identify abnormal fluctuations; Based on the heart rate sensor data, the user's heart rate variability index is evaluated, and the blood oxygen saturation data and body surface temperature data are combined to build a stress assessment model and output stress management suggestions; Based on the accelerometer and gyroscope data, combined with the height and weight information entered by the user, a deep learning model is built to identify the user's movement pattern and calculate the calories consumed; Provide personalized exercise plans and energy consumption analysis reports by analyzing the user's historical exercise habits and physical fitness level; Using time series analysis technology, track and analyze the user's heart rate and blood oxygen health indicators, predict future health trends, and discover potential health problems; Upload various health data to the cloud server for storage and backup.

2. The method for monitoring sports health of a smart watch according to claim 1, characterized in that: The time synchronization of continuous data and alignment of various monitoring indicators on the same time axis specifically includes: When the smartwatch is first started or the user changes the device, the time is calibrated and synchronized with the Internet time server through the Network Time Protocol; The smartwatch is connected to the smartphone via Bluetooth or Wi-Fi, obtains the timestamp of the smartphone, and sets it as the system time of the smartwatch; Each sensor records an initial timestamp when it is initialized, and uses it as the time reference for subsequent data collection; When the sensor collects a set of data, a timestamp is automatically added to the set of data based on the initial timestamp when the sensor was initialized and the current sampling interval; Establish a time synchronization mechanism. The time synchronization thread checks the data buffer of each sensor and aligns the data to the same time axis according to the timestamp; For heart rate data, linear interpolation is used to estimate the heart rate value between two sampling points; For the accelerometer data, resample it to reduce the data frequency so that it is aligned with the time axis of other sensor data; Dynamically adjust the sampling frequency of sensors based on the user's activity status and monitoring needs. When the user is exercising vigorously, increase the sampling frequency of the accelerometer and gyroscope. By comparing with the network time protocol server, the accuracy of the time synchronization mechanism can be verified.

3. The method for monitoring sports health of a smart watch according to claim 2, characterized in that: The use of machine learning algorithms to perform real-time analysis and anomaly detection on heart rate and blood oxygen saturation data and identify abnormal fluctuations therein specifically includes: The heart rate and blood oxygen saturation data were cleaned to remove values ​​beyond the physiological range. For missing values, the front and back value filling method was used to fill them. Standardize the data, convert it to the same order of magnitude, convert the data to a mean of 0 and a standard deviation in the range of [0,1]; Extracting derived features of the heart rate and blood oxygen saturation data, wherein the derived features include a rate of change of the heart rate and a fluctuation range of the blood oxygen saturation; Based on the characteristics of heart rate and blood oxygen saturation data, the long short-term memory network algorithm was selected as the machine learning algorithm for model training; Build a training dataset containing normal and abnormal heart rate and blood oxygen saturation data. Normal data comes from daily monitoring data of healthy users, and abnormal data comes from data of users with known health problems. The selected machine learning algorithm is trained using the training data set to obtain a model for detecting abnormal fluctuations in the user's heart rate and blood oxygen saturation data; The smart watch collects heart rate and blood oxygen saturation data in real time, forms a data stream, extracts features from the data stream, and obtains a real-time feature vector; The real-time feature vector is input into the abnormal fluctuation detection model of the user's heart rate and blood oxygen saturation data to obtain a prediction result, which is the probability of normal or abnormality of the user's current physiological state; Based on the prediction results, abnormal fluctuations in heart rate and blood oxygen saturation are identified and abnormal alarms are triggered. The abnormal alarms remind users through watch vibrations, screen displays, and mobile phone app notifications. Based on the accumulation of user data and the discovery of new physiological characteristics, the machine learning model is updated to adapt to changes in user physiological status and new monitoring needs.

4. The method for monitoring sports health of a smart watch according to claim 3, characterized in that: The method of evaluating the user's heart rate variability index based on the heart rate sensor data, and building a stress assessment model by integrating the blood oxygen saturation data and the body surface temperature data, and outputting stress management suggestions specifically includes: By obtaining the adjacent R wave intervals and N wave intervals in the heart rate data, the root mean square value of the difference between the average R wave interval and the adjacent NN interval is calculated, and the sum of the two values ​​is added to obtain the heart rate variability index; By performing Fourier transform on the heart rate variability signal and analyzing the power spectrum density of each frequency band, high frequencies are related to parasympathetic nerve activity, while low frequencies are related to sympathetic nerve activity; A heart rate variability index measurement is performed once when the user is still asleep to establish a personal baseline heart rate variability index value; Determine whether the real-time heart rate variability index value is lower than the baseline heart rate variability index value, if so, output that the user is in a high stress state, if not, no output; Determine whether the user's blood oxygen saturation is lower than a preset blood oxygen saturation threshold under high stress. If so, output that the user is in a hypoxic state. If not, no output is made. Determine whether the user's body surface temperature under high stress is higher than a preset body surface temperature threshold. If so, output that the user is in a high consumption state. If not, no output is made. Heart rate variability index, blood oxygen saturation and body surface temperature were selected as input variables of the stress assessment model, and the support vector machine method was used to build the stress assessment model to learn the relationship between the variables and the stress level; For users with high stress levels, meditation, deep breathing, and yoga are recommended to relieve stress and enhance parasympathetic nervous system activity; For users who are in hypoxic state, it is recommended that users perform moderate aerobic exercise; For users in a high-consumption state, it is recommended that they avoid overexertion, take rest, and adjust their breathing rhythm.

5. The method for monitoring sports health of a smart watch according to claim 4, characterized in that: The method of building a deep learning model based on the accelerometer and gyroscope data and the height and weight information input by the user to identify the user's exercise pattern and calculate the calories consumed specifically includes: Extracting features to be identified from accelerometer and gyroscope data and using them as input to a deep learning model, wherein the features to be identified include cadence, stride length, and rotation angle; A deep learning model for motion pattern recognition is built based on a convolutional neural network. The model structure includes convolutional layers, pooling layers, and fully connected layers. The convolutional layers are used to extract features, the pooling layers are used to reduce feature dimensions, and the fully connected layers are used for classification. Divide the preprocessed accelerometer and gyroscope data into training, validation, and test sets; The model is trained using the training set data, and the model parameters are updated through a back-propagation algorithm. The newly acquired accelerometer and gyroscope data are input into the trained model, and the model outputs the user's current motion mode, which includes walking, running, and cycling. Use the basal metabolic rate calculation formula and the exercise calorie calculation formula to calculate the total calorie consumption during exercise.

6. The method for monitoring sports health of a smart watch according to claim 5, characterized in that: The use of time series analysis technology to track and analyze the user's heart rate and blood oxygen health indicators, predict future health trends, and discover potential health problems specifically includes: Conduct visual analysis of the data, draw time series graphs, autocorrelation graphs, and partial autocorrelation graphs, and preliminarily determine the stability and trend of the data; The order of the autoregressive term is determined by observing the partial autocorrelation graph. The partial autocorrelation graph decays and approaches 0 after lagging the order. The optimal order value is selected by comparing the model fitting effects under different order values. The order of the moving average term is determined by observing the autocorrelation graph. The autocorrelation graph decays and approaches 0 after lagging the order. The optimal order value is selected by comparing the model fitting effects under different order values. Perform a stationarity test to determine whether the data is not stationary. If so, perform a difference process to convert it into a stationary sequence. The number of differences is determined by at least one difference and stationarity test. If not, no output is made. Based on the order of the autoregressive term, the number of differences and the order of the moving average term, an autoregressive integrated moving average model is constructed; Determine whether there is correlation between the residual sequences of the autoregressive integrated moving average model. If so, adjust the order of the model's autoregressive term, the number of differences, and the order of the moving average term. If not, no output is made. Using the trained model, the user's future heart rate and blood oxygen saturation are predicted in the short term, so that the user can understand his or her health status and take corresponding preventive measures; By analyzing long-term trends and cyclical changes in historical data, we can make long-term predictions about users' future health conditions, plan their health management strategies, and prevent potential health problems.

7. A sports health monitoring system for a smart watch, used to implement the sports health monitoring method for a smart watch as described in any one of claims 1 to 6, characterized in that: include: A data acquisition module, which is used to monitor the user's heart rate and blood oxygen saturation in real time through a heart rate sensor and a blood oxygen sensor, record the user's movement trajectory, steps, speed and distance information through an accelerometer and a gyroscope, detect environmental noise using a microphone, and monitor body surface temperature through a temperature sensor; A data processing module, which is used to synchronize the continuous data, align the monitoring indicators on the same time axis, and use machine learning algorithms to perform real-time analysis and anomaly detection on the heart rate and blood oxygen saturation data to identify abnormal fluctuations therein; Health monitoring module, which is used to evaluate the user's heart rate variability index based on heart rate sensor data, and to build a stress assessment model based on blood oxygen saturation data and body surface temperature data, output stress management suggestions, build a deep learning model based on accelerometer and gyroscope data, combined with the height and weight information input by the user to identify the user's exercise pattern, and calculate the calories consumed, and provide personalized exercise plans and energy consumption analysis reports by analyzing the user's historical exercise habits and physical fitness level; The data prediction module is used to use time series analysis technology to track and analyze the user's heart rate and blood oxygen health indicators, predict future health trends, discover potential health problems, and upload various health data to the cloud server for storage and backup.

Images